Examining Differences in the Individual and Contextual Risk Factors for Police Officer, Correctional Officer, and Non-Protective Service Suicides

Abstract

Suicide rates vary across occupational groups, with protective service occupations at elevated risk for suicide. Yet, research on correctional officer suicide remains sparse, as does research linking the broader social context to police officer suicides and correctional officer suicides. This study examines differences in the individual and contextual risk factors for police officer suicides, correctional officer suicides, and non-protective service suicides in the U.S. from 2003 to 2018 using data from the National Violent Death Reporting System and the American Community Survey. Descriptive analysis and logistic hierarchical models indicated that police officer suicides, correctional officer suicides, and non-protective service suicides had distinct etiologies. Additionally, the broader social context had an appreciable impact on occupation-specific suicide. Ultimately, interventions to address suicidality, including among public safety professionals, should be occupation-specific and consider both individual and contextual risk factors.

Keywords:

• Suicide
• law enforcement
• social context

Introduction

In 2020, approximately 46,000 persons died by suicide in the United States (Ehlman et al., 2022). Although this represents the second consecutive year-over-year decline of suicides following a peak in 2018, suicide remains a leading cause of death and a persistent public health problem in the U.S. Suicide has wide-reaching impact on families, significant others, friends, colleagues, and neighbors by triggering depression, anxiety, anger, guilt, and suicidal ideation (Chapman & Dixon-Gordon, 2007). Suicide also bears a significant financial burden: medical, employment leave-of-absence, and quality of life costs totaled nearly $500 billion in 2019 (Peterson et al., 2021).

The social burden of suicide is disproportionately distributed across occupational groups. In particular, the Centers for Disease Control and Prevention (CDC) indicate that protective service occupations are at an elevated risk for suicide. The protective services are among the ten major occupational groups with suicide rates higher than the general population for males. For females, protective service workers trail only construction and extraction workers for the highest suicide rates (Peterson et al., 2020). Public safety professions are dominated by precisely the demographic groups most affected by suicide—middle-aged, white males. Additionally, these professions are characterized by well-established occupational risk factors for suicide, including lower education, high job stress, and work-related access to lethal weapons (Milner et al., 2017). This recognition has inspired a renewed interest in law enforcement officer well-being and suicide, with local, state, and federal agencies identifying officer safety and health as strategic priorities (National Institute of Justice, 2018).

This study contributes to the literature by examining differences in the individual and contextual risk factors for police officer suicides, correctional officer suicides, and non-protective service suicides in the U.S. from 2003 to 2018. Data are drawn from the National Violent Death Reporting System (NVDRS) and the American Community Survey (ACS). Descriptive analysis and logistic hierarchical regression models are employed to investigate differences in the risk factors for suicide across occupational groups. The focus on correctional officers as distinct from police officers is motivated by the dearth of research on correctional officer suicide as well as by fundamental differences between these public service professions.

In part because they are not traditional “first responders,” correctional officers are generally excluded from studies of suicide among public safety personnel. For example, the most comprehensive systematic review of suicide among “our protectors” focused on “police officers, firefighters, emergency medical technicians (EMTs), and paramedics” (Stanley et al., 2016, p. 25). Correctional officers were conspicuously excluded. The omission may have been a product of the very few quantitative studies that have focused on correctional officer suicide. In one exception, Stack and Tsoudis (1997) found that the risk for suicide was 39% higher among correctional officers than among the general working age population in a multivariate logistic regression analysis of working population suicide across 21 states. Similarly, via a proportionate mortality ratio analysis, Violanti (2017) demonstrated that the risk for suicide was 41% higher among correctional officers than among suicide decedents aged 18–90 with a reported occupation or industry. Research also suggests that upward of 10% of correctional officers consider suicide (Lerman, 2017), while 1% actively plan their suicide (Spinaris & Brocato, 2019). High suicide rates contribute to the below-average life expectancy of correctional officers—59 years (Cheek, 1984). The results of these studies suggest that the context within which correctional officers work, and the nature of the work they do, are in some ways fundamentally different from those in other public safety occupations, including policing.

Police officers, firefighters, EMTs, and paramedics work directly with the public and typically perform their functions in public spaces. Much of their work involves interactions with law-abiding residents in the community. Correctional officers, on the other hand, work behind the walls of jails and prisons almost exclusively with individuals accused of and convicted of crime. Even though they provide a vital public safety function and deserve recognition as “protectors” (Stanley et al., 2016), they work within institutions and perform a far less “public” public safety function. This may be one of the most distinguishing features of correctional work.

Several decades ago, Farkas and Manning (1997) used qualitative methodologies to identify the similarities and differences between police work and correctional work. They share several key features, including interacting with people who are adversarial and oppositional. In addition, both occupations are “suffused with uncertainty and unpredictability,” as well as secrecy (Farkas & Manning, 1997, pp. 53–54). Notable though were the differences, including that “COs are bounded physically and locked in literally with those they ‘police’ or regulate,” and “their discretion is severely limited relative to the police officer” (Farkas & Manning, 1997, p. 62). Also notable, while trained in the use of and generally licensed to carry firearms, correctional officers are unarmed on the job (except for those responsible for perimeter security) and must rely on interpersonal skills to manage the populations they work with.

Additionally, the general public perceives corrections work as the least prestigious of the public safety occupations, ranking below police, firefighters, and EMTs (Sundt, 2009). Penologist Hans Toch made the point even more plainly: “Prison guards (correctional officers) are truly imprisoned: They are not only physically confined but are locked into movie caricatures, into pejorative prophecies (sometimes self-fulfilling), into anachronistic supervision patterns, into unfair civil service definitions, into undeserved hostilities and prejudgments of their actions. Officers are imprisoned by our ignorance of who they are and what they do, which is the price they pay for working behind walls” (quoted in Sundt, 2009, p. 40).

The unique nature of corrections and the environment in which correctional officers work lead to particularly high rates of mental health problems (Carleton et al., 2018; Easterbrook et al., 2022; St. Louis et al., 2023) and suicidal behavior (Frost & Monteiro, 2020; Ricciardelli et al., 2022). The leading social-psychological theory of suicide (Joiner, 2005; Van Orden et al., 2010) provides theoretical grounding for these findings. According to the interpersonal theory of suicide, a sense of thwarted belonging and perceived burdensomeness explain suicidal ideation. But, suicidal thoughts do not translate into action because most people do not have a capability for suicide (Joiner, 2005). Given high levels of exposure to violence, injury, and death (Spinaris et al., 2012), correctional officers engage in work that can lead to an acquired capability for suicide. They work in closed institutions where the potential for violence and the risk of personal injury are omnipresent (Ricciardelli et al., 2022). Correctional officers routinely have to break up fights and respond to serious assaults and self-harm incidents, including suicide attempts and suicides (Smith et al., 2019). Correctional officers also face threats to their own personal safety and that of their colleagues (Goulette et al., 2022). They have among the highest rates of intentional injury by another person and miss more days from work due to injury than any other category of government employee (Konda et al., 2012). Although merely a sentence in an article about occupational injuries, among those correctional officers fatally injured on the job between 1999 and 2008, more than one-third died by suicide (Konda et al., 2012).

Concern about correctional officer suicide has led several states to convene commissions, task forces, and surveys to address the problem. The Massachusetts legislature established a Special Commission to Study the Prevention of Suicide among Correction Officers (2019), California funded the Correctional Officer Survey (CCOS) on Health and Wellness (Lerman, 2017), the Michigan Department of Corrections partnered on a 2018–2019 survey to understand correctional staff well-being (Spinaris & Brocato, 2019), and New Jersey established a Task Force on Police Suicide (Milgram & Velez, 2009). From a policy perspective, there has been increased investment in understanding correctional officer suicide.

Taken together, the relative dearth of research on correctional officer suicide and the lack of comparative work suggest that we know very little about the potential impacts of occupational differences across policing and corrections. But, we know that differences exist, are theoretically meaningful, and are increasingly a focus of local, state, and federal agencies (National Institute of Justice, 2018). Below, we discuss the individual and contextual risk factors for suicide, with a focus on police officers and correctional officers.

Literature Review

Individual Risk Factors for Suicide

There are several individual factors associated with an increased risk for suicide. Regarding demographic characteristics, suicide rates are consistently highest among middle-aged, white males (Curtin & Hedegaard, 2019). This pattern holds true in policing (Krishnan et al., 2022) and corrections (Frost & Monteiro, 2020), professions dominated by these demographic groups (Milner et al., 2017).

Personal risk factors are among the most substantiated risk factors for suicide in the general population. Research has demonstrated that people who are single, divorced, or widowed are at increased risk of suicide compared with those who are married (Næss et al., 2021). The protective effects of marriage appear to be reinforced by having children and by marriage stability (Krishnan et al., 2022). Similarly, being married or in a committed relationship is a protective factor for police and correctional officers (Quarshie et al., 2021; Spinaris & Brocato, 2019; Violanti et al., 2009). Mental health problems, including depression, anxiety, and hopelessness, are common risk factors for suicide among the general population (Masango et al., 2008), police officers (Di Nota et al., 2020), and correctional officers (Frost & Monteiro, 2020). Research also suggests that differences in job-related physical violence and injury, job satisfaction, and perceived organizational support translate to higher rates of post-traumatic stress disorder (PTSD), depression, and anxiety among correctional officers, compared to other occupational groups (Regehr et al., 2021). Educational attainment is inversely associated with suicide risk in the general population (Abel & Kruger, 2005), and law enforcement professions are characterized by comparatively low levels of education (Milner et al., 2017). For example, being a correctional officer typically requires a high school diploma rather than a college degree (Butler et al., 2019). Substance use, and in particular alcohol use, is also strongly associated with suicidal behavior in the general population (Bilsen, 2018). Chae and Boyle (2013) reported that alcohol use and abuse is one of the five most prominent aspects of policing associated with suicide risk, and correctional officer suicide is often preceded by growing problems with substance use and abuse (Frost & Monteiro, 2020). Finally, direct and indirect exposure to violence are risk factors for suicide (Bilsen, 2018; Zimmerman, 2013), both of which are particularly common among police officers (Rouse et al., 2015) and correctional officers (Spinaris & Brocato, 2019; St. Louis et al., 2023).

Prior exposure to suicide and death are also strong predictors of subsequent suicide. In particular, there is a strong link between suicide, previous suicide attempts, and history of self-harm more generally (Cooper et al., 2005). Estimates indicate that between 25% and 50% of completed suicides are preceded by an earlier suicide attempt (Masango et al., 2008). Research on the association between completed suicide and previous suicide attempts among law enforcement officers is mixed. Some studies indicate that police officers who die by suicide are more likely to have a history of suicide attempts (Rouse et al., 2015), with many of these officers communicating their suicidal intent to others (Barron, 2010). Other studies suggest that law enforcement officer suicides are less likely than general population suicides to be accompanied by prior suicidal behavior or previous disclosure of suicidal intent (Milgram & Velez, 2009). Exposure to suicide and death are also risk factors for an “acquired capability for suicide” (Joiner, 2005, p. 47), and witnessing and hearing about suicide and death are endemic in policing and correctional work (Spinaris & Brocato, 2019).

Several life stressors have been linked to suicide risk. While the extent to which these stressors are precipitating factors for suicide varies by age group, financial strain (Elbogen et al., 2020), eviction and loss of housing (Fowler et al., 2015), job stress and job loss (Stack, 2021), physical health problems, and criminal and civil legal problems (Choi et al., 2022) are associated with suicide in the general population. Studies on police officer suicide have highlighted occupational stress due to burnout, in particular emotional exhaustion and depersonalization (Berg et al., 2003; Wray & Jarrett, 2019) and job dissatisfaction (Quarshie et al., 2021). Additionally, research has indicated that occupational stress is particularly acute in the correctional setting (Brower, 2013; Milgram & Velez, 2009). Studies suggest that correctional officers report significantly higher levels of organizational stress, especially related to staff and resource shortages and leadership issues, than police officers (Summerlin et al., 2010).

Relationship problems also precipitate suicidal behavior. Family conflict and intimate partner problems are strongly associated with suicide in both the general population (Choi et al., 2022) and among law enforcement personnel (Chae & Boyle, 2013). Additionally, aspects of policing and correctional work may exacerbate relationship problems. In particular, the strains of shift work (Milgram & Velez, 2009) and erratic shift schedules (Stanley et al., 2016) on family relationships are among the most prominent risk factors for law enforcement officer suicide.

Characteristics of the suicide, most notably suicide method, are also associated with suicide risk. Firearm access is a well-established risk factor for suicide (Miller et al., 2007), particularly among police (Stanley et al., 2016) and correctional officers (Milgram & Velez, 2009). Firearms are also more lethal than other suicide methods. Roughly 90% of suicide attempts with a firearm are fatal, compared to 56% for drowning, 53% for asphyxiation, 31% for poisoning by gas, and less than 1% for cutting (Conner et al., 2019). We also note that suicide may be more likely to be preceded by homicide among law enforcement officers. Exposure to violence, experiences with domestic violence, and access to firearms are correlates of homicide-suicide among law enforcement officers (Violanti, 2007).

Contextual Risk Factors for Suicide

Compared to research on the individual risk factors for suicide, studies on contextual risk factors for suicide are sparse. Much of the existing research is grounded in Durkheim’s (1951) sociological theory of suicide, first published in 1897, which contends that suicide can be explained in part by macro-level conditions such as social integration and regulation. For example, studies at the city, county, and national levels of analysis have demonstrated that contextual characteristics representing a lack of social integration and regulation (e.g. rates of marital instability, residential stability, divorce, non-marital births, and non-religiosity) are associated with comparatively high suicide rates (e.g. Baller & Richardson, 2002; O’Brien & Stockard, 2006). Research has also examined the effects of social disorganization on suicide rates. Theoretically, social disorganization may isolate individuals from private and public institutions of social control (Sampson et al., 2002; Shaw & McKay, 1942), thereby disrupting integration and regulation and leading to suicide (Maimon & Kuhl, 2008). Empirically, research has detected a positive, significant relationship between concentrated disadvantage and suicide rates (Kubrin et al., 2006; Wadsworth & Kubrin, 2007). Relatedly, longitudinal research has linked unemployment rates to suicide rates over time (Vigderhous & Fishman, 1978).

While research has investigated the link between contextual characteristics and suicide rates, few studies have examined the contextual risk factors for individual variation in suicide. The limited extant research has demonstrated that neighborhood residential mobility, household income, poverty, and racial concentration have a direct and positive effect on the risk of suicide, net of individual risk factors (Cubbin et al., 2000). Similarly, research has found that community religiosity is a protective factor against individual suicide (Van Tubergen et al., 2005), while community poverty increases the risk of suicidal behavior (Bernburg et al., 2009; Dupere et al., 2009). Research has also demonstrated that contextual factors condition the effects of individual risk factors on suicidal behavior. For example, Maimon and Kuhl (2008) found that the relationship between depression and youth attempted suicide was attenuated in communities with higher levels of religiosity; and Maimon et al. (2010) found that collective efficacy amplified the protective effects of parental attachment and support on youth attempted suicide.

The contextual risk factors for police officer and correctional officer suicide are generally overlooked in the literature. Given the widespread recognition that all forms of violence are best explained by an attendant focus on both individuals and social contexts (Krug et al., 2002), this study uses data from the NVDRS to examine the individual and contextual risk factors for police officer, correctional officer, and non-protective service suicide.

Methods

Data

The NVDRS is a state-based active web surveillance system of all violent deaths in the U.S. Established by the Centers for Disease Control and Prevention (CDC) in 2003, the NVDRS is a pooled cross-sectional time series, with cases nested within place and state of death. As of 2018, 41 states and the District of Columbia reported to the NVDRS. The majority of cases are from 17 states (see Appendix), but reported deaths occurred in all 50 states and the District of Columbia. The NVDRS integrates data from multiple sources, including death certificates (DC), coroner/medical examiner (CME) records, and law enforcement (LE) reports.

Using FIPS identifiers for the census place and state of death, we appended contextual information from the ACS, a continuous nationwide survey of 3.0 to 3.5 million households. Place and state data were derived from the five-year midpoint of the study period, 2009–2013.

Measures

Occupation-Specific Suicide

The key variable of interest is occupation-specific suicide, differentiating among police officer suicide, correctional officer suicide, and suicide in the general population. The NVDRS defines suicide as death resulting from the intentional use of force against oneself (ICD-10 codes X60–X84 and Y87.0). Detailed occupation information on suicide victims is provided from the 2010 Standard Occupation Classification (SOC). Police officers are defined as police and sheriff’s patrol officers and include transit police. Correctional officers include correctional officers and jailers and exclude bailiffs. Non-protective services represent all occupations exclusive of law enforcement workers, firefighters and prevention workers, and other protective service workers. The study sample includes 1,373 police officer suicides, 687 correctional officer suicides, and 199,838 non-protective service suicides in the U.S. from 2003 to 2018. Suicides were nested within 12,306 places and 50 states and the District of Columbia.

Individual Risk Factors for Suicide

Demographic characteristics

Demographic factors include age at time of death, biological sex, and race/ethnicity (white, black, Hispanic, and other).

Personal risk factors

Personal risk factors include marital status, educational attainment, mental health problems, alcohol and drug use, and involvement in interpersonal violence. Binary variables indicate whether or not (1 = yes; 0 = no) the victim was married, had been diagnosed with and/or treated for a mental health problem according to the DSM-V, had alcohol or illegal drugs present in their system at time of death, or had been a perpetrator or victim of interpersonal violence within the past month. Educational attainment was categorized as less than high school, high school degree or equivalent, or some college or higher.

Prior exposure to suicide and death

Binary measures of previous suicidal behavior include whether the individual had previously attempted suicide or disclosed suicidal intent to another person within the past month. Binary indicators of suicide exposure include whether the individual experienced a recent family member’s or friend’s suicide or death.

Life stressors

Binary indicators of life stressors include whether the individual faced criminal legal problems. This included impending arrest, evading law enforcement, or involvement in civil disputes (e.g. divorce, lawsuits, and custody battles). Measures also indicate whether the individual experienced physical health problems, faced job problems (e.g. unemployment or demotion), or experienced financial difficulties (e.g. bankruptcy, debt, or foreclosure).

Relationship risk factors

Binary measures of interpersonal conflict include whether the decedent had problems with an intimate partner, family member, or some other person. The analysis also includes an indicator of whether the suicide was precipitated by an argument.

Suicide characteristics

The analysis includes binary indicators of whether the suicide was completed by firearm, cutting with a sharp instrument, asphyxiation, poison (e.g. drug overdose, carbon monoxide inhalation), or another method (e.g. drowning, electrocution). Suicide location includes home, street, car, business, and other. Two additional binary indicators measure whether the suicide was part of a homicide-suicide or occurred during the commission of a crime.

Contextual Risk Factors for Suicide

This study includes measures of concentrated disadvantage, residential stability, and racial/ethnic heterogeneity at the place- and state-level, derived from the ACS. Concentrated disadvantage represents the weighted factor regression score of median household income (reverse-coded) and the proportion of the population aged 18–64 living below the poverty line, unemployed, single female-headed households with children, married (reverse-coded), and adults with a high school degree (reverse-coded) (Wikstrom & Loeber, 2000). Residential stability is the average of the proportion of the population living in owner-occupied housing and living in the same house for the past year (Sampson et al., 1997). Racial/ethnic heterogeneity was measured using Blau’s (1997) index: 1—∑p_i², where p_i is the proportion of the population in each racial/ethnic group. The analysis also controls for the natural log of the total population. Mardia’s (1970) tests for multivariate normality indicated that total population rejected normality for skewness and kurtosis. In contrast, the logged version of total population had a substantially lower chi-square value for skewness and was not significantly kurtotic.

Analytic Strategy

The analysis proceeds in two stages. The first stage of analysis examines whether there are differences in the risk factors for non-protective service suicides and law enforcement officer suicides. The second stage of analysis examines whether there are differences in the risk factors for police officer suicides and correctional officer suicides.

Chi-square tests, t-tests, and Wilcoxon-Mann Whitney tests provide an initial examination of the bivariate differences in risk factors across the two groups of interest. A three-level logistic regression model nesting suicides within place and state of death then examines the factors that distinguish the two groups of interest, controlling for the full array of study covariates. When comparing law enforcement officer suicides and non-protective service suicides, the dependent variable is coded zero for non-protective service suicides and unity for law enforcement officer suicides. When comparing law enforcement officer suicides and correctional officer suicides, the dependent variable is coded zero for police officer suicides and unity for correctional officer suicides. The multilevel models account for year of death via a series of dummy variables. All variance inflation factors (VIF) were below 2.55, with an average VIF of 1.28. Removing educational attainment—the only variable with VIF above 2.00—from the multilevel models generated identical findings with respect to the remaining variables, easing concerns about multicollinearity.

Results

Comparing Law Enforcement and Non-protective Service Suicides

Table 1 provides a descriptive comparison of differences in the risk factors for suicide across non-protective service suicides and law enforcement officer suicides. Regarding demographic characteristics, law enforcement officers who died by suicide were older and less likely to be female and non-white than non-protective service workers who died by suicide.

Table 1. Descriptive statistics for non-protective service suicides and law enforcement suicides, NVDRS, 2003–2018, N = 199,838 non-protective service suicides, 2,060 law enforcement suicides, 12,306 U.S. cities, 50 U.S. states and the District of Columbia.

Variable	Non-Protective Services (n = 199,838)			Law Enforcement (n = 2,060)
	% / Mean	N / (SD)	[Range]	% / Mean	N / (SD)	[Range]
Individual Risk Factors for Suicide
Demographic Characteristics
Age***
Under 30	44,815	22.42%		208	10.10%
30–39	32,808	16.42%		353	17.14%
40–49	37,693	18.86%		449	21.80%
50–59	39,263	19.65%		389	18.88%
60 and older	45,259	22.65%		661	32.08%
Sex***
Female	22.68%	45,330		9.71%	200
Male	77.32%	154,508		90.29%	1,860
Race/Ethnicity***
White	82.75%	165,346		84.08%	1,732
Black	6.73%	13,453		8.77%	181
Hispanic	5.87%	11,734		4.57%	94
Other	4.65%	9,305		2.58%	53
Personal Risk Factors
Marital Status***
Married	32.84%	65,631		50.18%	1,034
Not Married	67.16%	134,207		49.82%	1,026
Educational Attainment***
Less than High School	18.57%	37,102		5.40%	111
High School	39.55%	79,041		37.53%	773
Some College or Higher	41.88%	83,695		57.07%	1,176
Mental Health Problems***	48.57%	97,062		38.56%	794
Alcohol in System at Death	33.58%	67,109		33.93%	699
Drugs in System at Death***	30.09%	60,128		20.67%	426
Recent Violence Perpetration**	3.09%	6,174		4.26%	88
Recent Victimization	.49%	985		.26%	5
Prior Exposure to Suicide and Death
Previous Suicide Attempt***	20.30%	40,558		10.55%	217
Disclosed Suicidal Intent	26.30%	52,548		25.42%	524
Recent Family/Friend Suicide	2.20%	4,402		1.94%	40
Recent Family/Friend Death	6.44%	12,870		6.22%	128
Life Stressors
Criminal Legal Problems	9.14%	18,258		8.36%	172
Civil Legal Problems*	3.57%	7,130		4.54%	94
Health Problems***	21.50%	42,975		29.17%	601
Job Problems***	11.16%	22,310		15.21%	314
Money Problems	10.19%	20,369		9.33%	192
Relationship Risk Factors
Intimate Partner Problems***	28.77%	57,486		36.99%	762
Family Problems**	7.67%	15,324		5.89%	121
Other Relationship Problems**	5.09%	10,164		3.57%	74
Argument Led to Suicide	11.51%	23,003		11.95%	246
Suicide Characteristics
Suicide Method***
Firearm	49.39%	98,699		80.07%	1,649
Cutting	2.04%	4,077		.59%	12
Asphyxiation	27.25%	54,461		10.92%	225
Poison	15.60%	31,178		6.88%	142
Other	5.72%	11,423		1.54%	32
Suicide Location
Home	74.63%	149,125		75.23%	1,550
Street	4.48%	8,946		4.99%	103
Car	4.49%	8,982		7.21%	149
Business	3.78%	7,557		3.17%	64
Other	12.62%	25,228		9.40%	194
Part of Homicide-Suicide***	1.70%	3,404		2.77%	57
During Crime Commission*	2.70%	5,403		3.53%	73
Year*
2003	1.43%	2,866		1.65%	34
2004	3.35%	6,686		3.45%	71
2005	3.84%	7,673		2.96%	61
2006	3.92%	7,843		3.50%	72
2007	4.09%	8,183		3.93%	81
2008	4.12%	8,228		4.08%	84
2009	4.29%	8,582		3.83%	79
2010	4.44%	8,870		4.61%	95
2011	5.23%	10,452		5.29%	109
2012	5.40%	10,785		5.63%	116
2013	5.51%	11,009		5.53%	114
2014	6.28%	12,553		5.53%	114
2015	8.82%	17,631		8.74%	180
2016	11.17%	22,331		13.06%	269
2017	12.82%	25,629		12.82%	264
2018	15.29%	30,517		15.39%	317
Contextual Risk Factors for Suicide
City-Level Characteristics
Population***	1.42	(1.25)	[-3.67–4.81]	1.22	(1.21)	[-2.33–4.35]
Disadvantage	-.16	(.85)	[-6.71–4.25]	-.13	(.87)	[-3.37–2.61]
Residential Stability***	-.62	(.91)	[-5.71–2.25]	-.52	(.93)	[-4.97–2.25]
Racial/Ethnic Heterogeneity**	.83	(1.00)	[-1.33–2.64]	.76	(1.01)	[-1.33–2.64]
State-Level Characteristics
Population***	.50	(.61)	[-1.80–2.23]	.56	(.59)	[-1.71–2.23]
Disadvantage**	.14	(.84)	[-1.72–2.49]	.09	(.83)	[-1.72–1.60]
Residential Stability*	-.02	(.68)	[-4.23–1.58]	.01	(.67)	[-2.47–1.58]
Racial/Ethnic Heterogeneity***	.25	(.76)	[-2.01–1.63]	.31	(.76)	[-2.01–1.63]

ABBREVIATION: SD = standard deviation.

*p < .05; **p < .01; ***p < .001 (two-tailed tests). Indicates statistically significant difference between non-protective service suicides and law enforcement suicides.

With respect to personal risk factors, law enforcement officers who died by suicide were more likely to be married, educated, and to have perpetrated violence than non-protective service workers, but less likely than non-protective service workers to have mental health problems and to have drugs in their system at the time of death. Regarding the prior exposure to suicide variables, law enforcement officers were less likely than non-protective service workers to have a known history of attempted suicide. Of the life stressors, law enforcement officers were more likely than non-protective service workers to have civil legal problems, physical health problems, and job problems. Of the relationship variables, law enforcement officers were more likely than non-protective service workers to have intimate partner problems, but less likely to have family problems or other relationship problems. Regarding the suicide characteristics, law enforcement officers were more likely than non-protective service workers to use a firearm as suicide method, to commit homicide-suicide, and to die by suicide during the commission of a crime.

With respect to the contextual risk factors, law enforcement suicides occurred in less populous, more stable, and less racially/ethnically heterogeneous places than non-protective service worker suicides. At the state level, law enforcement suicides occurred in more populous, less disadvantaged, more stable, and more racially/ethnically heterogeneous states.

Table 2 examines the risk factors that distinguish law enforcement officer and non-protective service worker suicides, controlling for the full array of study covariates and accounting for clustering within places and states. The results confirm the majority of findings from the descriptive analysis. Regarding demographic characteristics, law enforcement officers were older (OR = 1.01, 95% CI = 1.01,1.01), less likely to be female (OR=.55, 95% CI=.51,.60), and more likely to be black (OR = 1.57, 95% CI = 1.41,1.74) than non-protective service workers.

Table 2. Three-level logistic regression model examining the individual and contextual factors that distinguish non-protective service suicides (zero) and law enforcement officer suicides (unity), NVDRS, 2003–2018, N = 199,838 non-protective service suicides and 2,060 law enforcement suicides, 12,306 U.S. cities, 50 U.S. states and the District of Columbia.^a

Variable	OR	95% CI
Individual Risk Factors for Suicide
Demographic Characteristics
Age	1.01***	[1.01, 1.01]
Sex (Female)^b	.55***	[.51, .60]
Race/Ethnicity^c
Black	1.57***	[1.41, 1.74]
Hispanic	1.14	[.99, 1.30]
Personal Risk Factors
Marital Status (Married)^d	1.47***	[1.39, 1.55]
Educational Attainment^e
Less than High School	.26***	[.22, .31]
High School	.66***	[.62, .71]
Mental Health Problems	.86***	[.81, .91]
Alcohol in System at Death	1.00	[.93, 1.08]
Drugs in System at Death	.79***	[.71, .87]
Recent Violence Perpetration	.99	[.84, 1.17]
Recent Victimization	.89	[.52, 1.51]
Prior Exposure to Suicide and Death
Previous Suicide Attempt	.79***	[.72, .86]
Disclosed Suicidal Intent	1.00	[.92, 1.07]
Recent Family/Friend Suicide	1.08	[.87, 1.34]
Recent Family/Friend Death	1.05	[.93, 1.18]
Life Stressors
Criminal Legal Problems	1.01	[.89, 1.14]
Civil Legal Problems	1.19*	[1.03, 1.37]
Health Problems	1.21***	[1.12, 1.31]
Job Problems	1.40***	[1.28, 1.53]
Money Problems	.75***	[.67, .83]
Relationship Risk Factors
Intimate Partner Problems	1.53***	[1.42, 1.65]
Family Problems	.97	[.85, 1.10]
Other Relationship Problems	.94	[.80, 1.10]
Argument Led to Suicide	.87**	[.79, .97]
Suicide Characteristics
Suicide Method^f
Cutting	.19***	[.13, .28]
Asphyxiation	.31***	[.28, .33]
Poison	.40***	[.36, .44]
Suicide Location^g
Street	1.05	[.91, 1.21]
Car	1.28***	[1.14, 1.44]
Business	.99	[.85, 1.15]
Part of Homicide-Suicide	.87	[.71, 1.06]
During Crime Commission	.97	[.81, 1.16]
Contextual Risk Factors for Suicide
City-Level Characteristics
Population	.94*	[.89, .99]
Disadvantage	.92**	[.86, .98]
Residential Stability	1.04	[.98, 1.11]
Racial/Ethnic Heterogeneity	1.03	[.97, 1.10]
State-Level Characteristics
Population	1.27**	[1.09, 1.48]
Disadvantage	.80**	[.70, .92]
Residential Stability	.97	[.83, 1.14]
Racial/Ethnic Heterogeneity	1.11	[.95, 1.31]

ABBREVIATION: SE = standard error.

^a Model includes dummy variables for year of death from 2003 to 2017 (reference = 2018).

^b Reference = male.

^c Reference = white. Model also controls for “other” race/ethnicity.

^d Reference = not married.

^e Reference = some college or higher.

^f Reference = firearm. Model also controls for “other” suicide methods.

^g Reference = home. Model also controls for “other” suicide location.

* p < .05.

** p < .01.

*** p < .001 (two-tailed test).

With respect to personal risk factors, law enforcement officers were more likely to be married (OR = 1.47, 95% CI = 1.39,1.55) and to have a college education (OR=.26, 95% CI=.22,.31 for less than high school; OR=.66, 95% CI=.62,.71 for high school) than non-protective service workers, but less likely to have mental health problems (OR=.86, 95% CI=.81,.91) and to have drugs in their system at time of death (OR=.79, 95% CI=.71,.87) than the general population. Regarding prior exposure to suicide, law enforcement officers were less likely to have a history of suicide attempt (OR=.79, 95% CI=.72,.86).

Of the life stressors and relationship factors, law enforcement officers were more likely than non-protective service workers to have civil legal problems (OR = 1.19, 95% CI = 1.03,1.37), health problems (OR = 1.21, 95% CI = 1.12,1.31), job problems (OR = 1.40, 95% CI = 1.28,1.53), and intimate partner problems (OR = 1.53, 95% CI = 1.42,1.65), but less likely to have financial problems (OR=.75, 95% CI=.67,.83) and triggering arguments (OR=.87, 95% CI=.79,.97). Regarding suicide characteristics, law enforcement officers were more likely than the general population to use a firearm, relative to cutting (OR=.19, 95% CI=.13,.28), asphyxiation (OR=.31, 95% CI=.28,.33), or poisoning (OR=.40, 95% CI=.36,.44). Officers were also more likely to die by suicide in a car, compared to at home (OR = 1.28, 95% CI = 1.14,1.44). Finally, law enforcement officers died by suicide in less disadvantaged places (OR=.92, 95% CI=.86,.98) and states (OR=.80, 95% CI=.70,.92) than non-protective service workers.

Comparing Police Officer and Correctional Officer Suicides

Table 3 provide a descriptive comparison of police officer and correctional officer suicides. Regarding demographic characteristics, correctional officers who died by suicide were younger and more likely to be female and non-white than police officers who died by suicide.

Table 3. Descriptive statistics for police officer suicides and correctional officer suicides, NVDRS, 2003–2018, N = 1,373 police officer suicides, 687 correctional officer suicides, 1,372 U.S. cities, 40 U.S. states.

Variable	Police Officers (n = 1,373)			Correctional Officers (n = 687)
	% / Mean	N / (SD)	[Range]	% / Mean	N / (SD)	[Range]
Individual Risk Factors for Suicide
Demographic Characteristics
Age***
Under 30	103	7.50%		105	15.28%
30–39	223	16.24%		130	18.92%
40–49	275	20.03%		174	25.34%
50–59	250	18.21%		139	20.23%
60 and older	522	38.02%		139	20.23%
Sex***
Female	7.65%	105		13.83%	95
Male	92.35%	1,268		86.17%	592
Race/Ethnicity***
White	86.54%	1,188		79.18%	544
Black	6.82%	94		12.66%	87
Hispanic	3.87%	53		5.97%	41
Other	2.77%	38		2.19%	15
Personal Risk Factors
Marital Status
Married	51.70%	710		47.15%	324
Not Married	48.30%	663		52.85%	363
Educational Attainment***
Less than High School	4.49%	62		7.22%	50
High School	32.84%	451		46.92%	322
Some College or Higher	62.67%	860		45.86%	315
Mental Health Problems	37.87%	520		39.96%	275
Alcohol in System at Death	34.66%	476		32.46%	223
Drugs in System at Death	19.93%	274		22.14%	152
Recent Violence Perpetration	3.63%	50		5.52%	38
Recent Victimization	.31%	4		.16%	1
Prior Exposure to Suicide and Death
Previous Suicide Attempt*	9.33%	128		13.00%	89
Disclosed Suicidal Intent	24.91%	342		26.43%	182
Recent Family/Friend Suicide	1.80%	25		2.23%	15
Recent Family/Friend Death	6.12%	84		6.42%	44
Life Stressors
Criminal Legal Problems*	7.37%	101		10.33%	71
Civil Legal Problems	4.48%	62		4.67%	32
Health Problems***	33.50%	460		20.52%	141
Job Problems	15.20%	209		15.25%	105
Money Problems	8.94%	123		10.12%	70
Relationship Risk Factors
Intimate Partner Problems***	34.22%	470		42.52%	292
Family Problems*	4.97%	68		7.73%	53
Other Relationship Problems	3.80%	52		3.11%	21
Argument Led to Suicide	11.16%	153		13.54%	93
Suicide Characteristics
Suicide Method***
Firearm	83.58%	1,147		73.05%	501
Cutting	.41%	6		.95%	7
Asphyxiation	8.91%	122		14.92%	103
Poison	5.58%	77		9.49%	65
Other	1.52%	21		1.59%	11
Suicide Location
Home	75.44%	1,035		74.82%	514
Street	4.62%	64		5.72%	39
Car	7.18%	99		7.26%	50
Business	3.12%	43		3.28%	23
Other	9.64%	132		8.92%	61
Part of Homicide-Suicide	2.55%	35		3.20%	22
During Crime Commission	3.30%	45		3.99%	27
Year***
2003	1.75%	24		1.46%	10
2004	3.13%	43		4.08%	28
2005	2.62%	36		3.64%	25
2006	3.64%	50		3.20%	22
2007	4.44%	61		2.91%	20
2008	4.15%	57		3.93%	27
2009	3.86%	53		3.78%	26
2010	5.32%	73		3.20%	22
2011	5.17%	71		5.53%	38
2012	5.83%	80		5.24%	36
2013	4.88%	67		6.84%	47
2014	5.39%	74		5.82%	40
2015	8.74%	120		8.73%	60
2016	13.18%	181		12.81%	88
2017	11.87%	163		14.70%	101
2018	16.03%	220		14.12%	97
Contextual Risk Factors for Suicide
City-Level Characteristics
Population***	.47	(1.15)	[-3.01–3.38]	.23	(1.14)	[-2.39–3.39]
Disadvantage***	.04	(.96)	[-3.59–2.67]	-.23	(.92)	[-2.98–2.94]
Residential Stability*	-.12	(.99)	[-4.82–2.68]	-.22	(.93)	[-3.12–2.74]
Racial/Ethnic Heterogeneity*	.27	(1.01)	[-1.85–2.12]	.16	(1.01)	[-1.86–2.12]
State-Level Characteristics
Population	.49	(.68)	[-1.94–2.40]	.50	(.66)	[-2.09–2.40]
Disadvantage*	.12	(.92)	[-1.84–1.83]	.23	(.91)	[-1.84–1.83]
Residential Stability	-.09	(.75)	[-2.90–1.60]	-.11	(.76)	[-2.90–1.68]
Racial/Ethnic Heterogeneity*	.31	(.75)	[-2.05–1.60]	.22	(.78)	[-2.05–1.60]

ABBREVIATION: SD = standard deviation.

*p < .05; **p < .01; ***p < .001 (two-tailed tests). Indicates statistically significant difference between police officer suicides and correctional officer suicides.

With respect to personal risk factors, correctional officers were less likely to have at least some college education than police officers. Regarding prior exposure to suicide, correctional officers were more likely than police officers to have a history of suicide attempt. Of the life stressors and relationship factors, correctional officers were more likely than police officers to have criminal legal problems, intimate partner problems, and family problems, but less likely to have physical health problems. Regarding suicide characteristics, correctional officers were less likely than police officers to use a firearm. Finally, correctional officer suicides occurred in less populous, less disadvantaged, less stable, and less racially/ethnically heterogeneous places than police officer suicides. At the state-level, correctional officer suicides occurred in more disadvantaged but less heterogeneous states than police officer suicides.

Table 4 examines the risk factors that distinguish police officer and correctional officer suicides, controlling for the study covariates. Regarding the demographic characteristics, correctional officers were younger (OR=.97, 95% CI=.96,.98) and more likely to be female (OR = 1.73, 95% CI = 1.25,2.39) and black (OR = 2.02, 95% CI = 1.40,2.91) than police officers. With respect to personal risk factors, correctional officers were less likely than police officers to have a college education (OR = 2.80, 95% CI = 1.48,5.29 for less than high school; OR = 2.21, 95% CI = 1.74,2.81 for high school) and to have alcohol in their system at time of death (OR=.73, 95% CI=.56,.97). Of the suicide characteristics, correctional officers were less likely than police officers to use a firearm, relative to asphyxiation (OR = 1.68, 95% CI = 1.22,2.32) and poisoning (OR = 1.81, 95% CI = 1.22,2.68). Regarding the contextual risk factors, correctional officers died by suicide in less disadvantaged cities (OR=.73, 95% CI=.62,.86) and less racially/ethnically heterogeneous states (OR=.74, 95% CI=.59,.92) than police officers.

Table 4. Three-level logistic regression model examining the individual and contextual factors that distinguish police officer suicides (zero) and correctional officer suicides (unity), NVDRS, 2003–2018, N = 1,373 police officer suicides, 687 correctional officer suicides, 1,372 U.S. cities, 40 U.S. states.^a

Variable	OR	95% CI
Individual Risk Factors for Suicide
Demographic Characteristics
Age	.97***	[.96, .98]
Sex (Female)^b	1.73***	[1.25, 2.39]
Race/Ethnicity^c
Black	2.02***	[1.40, 2.91]
Hispanic	1.50	[.94, 2.38]
Personal Risk Factors
Marital Status (Married)^d	.99	[.81, 1.22]
Educational Attainment^e
Less than High School	2.80**	[1.48, 5.29]
High School	2.21***	[1.74, 2.81]
Mental Health Problems	.99	[.80, 1.23]
Alcohol in System at Death	.73*	[.56, .97]
Drugs in System at Death	.99	[.73, 1.36]
Recent Violence Perpetration	1.35	[.72, 2.55]
Recent Victimization	.60	[.08, 4.76]
Prior Exposure to Suicide and Death
Previous Suicide Attempt	1.18	[.84, 1.66]
Disclosed Suicidal Intent	1.11	[.85, 1.45]
Recent Family/Friend Suicide	1.21	[.58, 2.51]
Recent Family/Friend Death	1.17	[.75, 1.82]
Life Stressors
Criminal Legal Problems	1.05	[.71, 1.55]
Civil Legal Problems	.79	[.48, 1.29]
Health Problems	.79	[.60, 1.05]
Job Problems	.76	[.56, 1.04]
Money Problems	1.08	[.74, 1.59]
Relationship Risk Factors
Intimate Partner Problems	1.09	[.85, 1.39]
Family Problems	1.42	[.90, 2.25]
Other Relationship Problems	.73	[.41, 1.32]
Argument Led to Suicide	.96	[.68, 1.35]
Suicide Characteristics
Suicide Method^f
Cutting	3.24	[.89, 11.79]
Asphyxiation	1.68***	[1.22, 2.32]
Poison	1.81**	[1.22, 2.68]
Suicide Location^g
Street	.96	[.60, 1.53]
Car	.84	[.57, 1.25]
Business	.72	[.40, 1.28]
Part of Homicide-Suicide	.94	[.45, 1.98]
During Crime Commission	.77	[.40, 1.49]
Contextual Risk Factors for Suicide
City-Level Characteristics
Population	.82**	[.71, .94]
Disadvantage	.73***	[.62, .86]
Residential Stability	1.00	[.85, 1.18]
Racial/Ethnic Heterogeneity	.91	[.76, 1.08]
State-Level Characteristics
Population	1.04	[.84, 1.29]
Disadvantage	1.12	[.95, 1.32]
Residential Stability	.83	[.69, 1.00]
Racial/Ethnic Heterogeneity	.74*	[.59, .92]

ABBREVIATION: SE = standard error.

^a Model includes dummy variables for year of death from 2003 to 2017 (reference = 2018).

^b Reference = male.

^c Reference = white. Model also controls for “other” race/ethnicity.

^d Reference = not married.

^e Reference = some college or higher.

^f Reference = firearm. Model also controls for “other” suicide methods.

^g Reference = home. Model also controls for “other” suicide location.

* p < .05.

** p < .01.

*** p < .001 (two-tailed test).

Discussion

Occupational health data suggest that suicide rates are elevated among public safety professionals (Violanti et al., 2013; Violanti & Steege, 2021). The findings herein correspond with research suggesting that occupational risk factors for suicide are inherent in law enforcement, including job problems, physical health problems, intimate partner problems, and civil legal problems (Peterson et al., 2020). The findings also indicated that law enforcement officers were more likely than the general population to use a firearm as suicide method. This finding is consistent with research substantiating access to firearms as a risk factor for suicide (Milner et al., 2017) and studies indicating that more than 80% of law enforcement officers who die by suicide use a firearm, most times a work-issued firearm (Chae & Boyle, 2013).

There are several factors that are relevant to both police officer and correctional officer suicide, including exposure to violence and trauma, shift work, and lack of organizational support (Wills et al., 2021). But, the analysis revealed several key differences between police officer and correctional officer suicides. In particular, correctional officers were less likely than police officers to have a college education, to have alcohol in their system at time of death, and to use a firearm as suicide method. These findings are consistent with the growing awareness that correctional officers may be differentially exposed to risk factors for suicide, including lower levels of educational attainment, more extensive histories of suicidal behavior, more criminal legal problems, and more intimate partner and family problems. Moreover, studies have demonstrated that correctional officers have suicide rates exceeding those of police officers (Brower, 2013; Milgram & Velez, 2009).

The analysis also demonstrated that contextual characteristics differentiated suicide across occupational groups. Law enforcement officers died by suicide in less disadvantaged places and states than non-protective service workers, and correctional officers died by suicide in less disadvantaged contexts than police officers. Broadly speaking, these findings reaffirm the fundamental premise that social-psychological behaviors cannot be disentangled from the broader social contexts in which they occur (Bronfenbrenner, 1994). Thus, suicide is best explained by an attendant focus on both individual and contextual factors (Krug et al., 2002; Fridel & Zimmerman, 2019).

The findings and conclusions may be tempered by several data limitations. First, the majority of suicide deaths in the study sample occurred in 17 states (see Appendix), representing 76.42% (N = 154,291) of the suicides in the study sample. These states are not equally dispersed across the four regions of the country. The South is overrepresented in the sample (including Georgia, Kentucky, Maryland, North Carolina, Oklahoma, South Carolina, and Virginia), while the Midwest is underrepresented (with only Ohio and Wisconsin). Additionally, the sample is skewed toward more densely populated states, which may be problematic as certain risk factors for suicide (e.g. depression, social isolation, and access to firearms) are more prevalent in rural communities than in urban communities (Hirsch, 2006). The unbalanced data across region and urbanicity limits generalizability and introduces the potential for bias. As such, the estimates in this study may be artificially inflated or attenuated and should be interpreted with caution.

However, prior research suggests that individual risk factors for suicide (e.g. history of depression and other mood disorders) are remarkably consistent across space, even across nations (Nock et al., 2008). Furthermore, while cultural and regional differences may impact rates of suicide, there is little evidence to suggest that they systematically affect suicides by occupation. That is, such biases should be consistent across police officer suicides, correctional officer suicides, and non-protective service worker suicides, thereby having minimal impact on the multivariate logistic regression results. To investigate this assertion, we re-estimated the statistical models after controlling for region of the country at the state level (Northeast, South, Midwest, and West). Regional differences were not detected when comparing police officer suicides and correctional officer suicides. The only significant difference when comparing non-protective service suicides and law enforcement officer suicides pertained to the Northeast region of the United States versus the Southern region of the United States. Compared to non-protective service suicides, law enforcement officer suicides were more likely to occur in the Northeast than in the South. Importantly, inclusion of the dummy variables for region did not alter the findings with respect to the other study covariates. Controlling for the region of the country in which suicides occurred, law enforcement officers died by suicide in less disadvantaged contexts than non-protective service workers, and correctional officers died by suicide in less disadvantaged contexts than police officers. Nevertheless, as the NVDRS releases nationally representative data, future research should replicate these findings to determine if they are generalizable to the nation as a whole.

Second, as the NVDRS only reports current occupation, any protective service workers who were unemployed or retired at the time of their death would be incorrectly classified (likely as non-protective service workers). While the true number of these false negatives is impossible to estimate, prior research on both police and correctional officers suggests that the majority of protective service worker suicides occur during employment. For example, research indicates that suicide rates are 8.4 times higher for working police officers than retired or separated officers (Violanti et al., 2011). Additionally, studies suggest that two-thirds of police officers and three-fourths of correctional officers are currently employed at the time of their suicide (Frost & Monteiro, 2020; Violanti et al., 2011). Nonetheless, the misclassification of retired and unemployed law enforcement officers may introduce bias into the results of analyses comparing law enforcement officer suicides and non-protective service worker suicides (Table 2). In particular, retired law enforcement officers are older and more likely to suffer from physical health problems, while unemployed law enforcement officers are more likely to suffer from job and money problems than their working counterparts. Given that the odds ratios for age, physical health problems, and job problems were significant and positive, these estimates are likely to be conservative. The study findings regarding money problems, however, may be biased down by misclassification of former officers as non-protective service workers, and should be interpreted with caution. With respect to the comparison of police officer suicides and correctional officer suicides, there is no reason to believe that the NVDRS systematically misclassifies retired or unemployed police officers more or less often than correctional officers. As a result, the analysis comparing the individual and contextual risk factors of police and correctional officer suicides are unlikely to be impacted by the NVDRS occupational coding protocol.

Third, the analysis includes place- and state-level data from the American Community Survey (ACS), derived from the five-year midpoint of the study period, 2009–2013. We acknowledge that the study period covers three full five-year cycles of the ACS (2005–2009, 2009–2013, and 2014–2018), and utilizing a single five-year period to capture sixteen years of data may be problematic given potential variation in contextual factors over time. However, the NVDRS is a cross-sectional panel that includes information on one time period per individual (the year of death), rather than a longitudinal time series. As such, any analysis focusing on individual-level differences, as herein, necessitates a cross-sectional approach, and the only way to model variation in context over time would be to examine counts or rates of suicides by place and state. While this analytical strategy could be used to address interesting research questions, it would omit rich individual-level data pertinent to individual variation in suicide. Moreover, the current study focuses on both individual and contextual risk factors, necessitating a multilevel modeling strategy and precluding an aggregate, longitudinal modeling strategy. Nevertheless, we examined variation in the contextual measures over the study period (for the three available five-year time periods). The lowest inter-time period correlation at the place-level was .99 for population, .80 for concentrated disadvantage, .70 for residential stability, and .84 for racial/ethnic heterogeneity. This suggests that the contextual characteristics remained largely consistent over time, providing justification for our measurement and analytical strategy.

Finally, because the NVDRS is a victim-based dataset, we were unable to include comparison groups of persons who did not die by suicide. While beyond the scope of this study, including a comparison group of persons who did not die by suicide would allow for an examination of the factors that can increase the risk for suicide or protect against it. Existing data sources such as the Mortality Disparities in American Community Study (MDAC) would be helpful in this regard.

With these limitations in mind, the findings suggest that suicide interventions should address the unique risk factors for law enforcement officer suicide, including job problems, physical health problems, intimate partner problems, civil legal problems, and access to firearms. To better understand the nature and extent of law enforcement officer suicide, the Federal Bureau of Investigation (FBI) launched the Law Enforcement Suicide Data Collection (LESDC) Act in January 2022 to compile national statistics on law enforcement suicide (Perine, 2021). Such national efforts, when coupled with published research on police officer (Chae & Boyle, 2013) and correctional officer (Frost & Monteiro, 2020) suicide, can facilitate effective public health resource allocation and interventions.

Relatedly, the findings demonstrated that law enforcement officers who died by suicide were less likely to have mental health problems than non-protective workers who died by suicide, but we suspect that this finding represents officer reluctancy to seek care for mental health-related concerns. For example, a nationwide study of American law enforcement personnel demonstrated that over 90% of officers perceived mental health stigma as one of the most significant barriers to help-seeking behavior (Drew & Martin, 2021; Wills et al., 2021). This stigma is associated with law enforcement cultural ideologies and acceptance of day-to-day occupational stress (Velazquez & Hernandez, 2019), and interacts with concerns about confidentiality to further reduce help-seeking behavior (Newell et al., 2022; Wills et al., 2021).

Importantly, if stigma remains, mental health and suicide intervention efforts “will be rendered ineffective, or at the very least, severely limited” (Drew & Martin, 2021, p. 295). Thus, effective interventions require addressing stigma and barriers to help-seeking behavior. There is optimism in this regard. For example, police forces in the United Kingdom have utilized Trauma Risk Management (TRiM)—an intervention aimed at reducing stigma and encouraging help-seeking behavior—with some success. A questionnaire of three police forces using TRiM and two police forces without TRiM indicated that the TRiM group reported less stigmatized views toward mental health issues and perceived fewer barriers to help-seeking (Watson & Andrews, 2018). Similarly, the Road to Mental Readiness (R2MR) program in Canada demonstrated modest reductions in stigma toward help-seeking among police (Carleton et al., 2018). And, the U.S. Bureau of Justice Assistance (BJA) anticipates the launch of a correctional officer and staff safety and wellness center that will serve as a national repository for best practices and evidence-based programs, including suicide awareness and prevention.

We conclude by reiterating the key insight that suicide interventions, like all violence prevention efforts, should be multisystemic and interdisciplinary. Advances in this regard include the National Consortium on Preventing Law Enforcement Suicide, convened by the International Association of Chiefs of Police (IACP), which takes a multidisciplinary approach to preventing officer suicide by engaging law enforcement professionals, mental health and suicide prevention experts, academics, families, and community stakeholders. Similarly, through the BJA, the National Suicide Awareness for Law Enforcement Officers (SAFLEO) Program takes a multifaceted approach to raising awareness, reducing stigma, and preventing officer suicide by engaging law enforcement agencies and staff, families, and communities. Ultimately, addressing officer deaths from suicide, which surpass officer deaths due to on-duty accidents and felonious assaults (Heyman et al., 2018; Violanti & Steege, 2021), requires comprehensive, multidisciplinary-informed interventions.

Appendix

State participation in the national violent death reporting system (NVDRS) by year, 2003–2018

Table

State	2003	2004	2005	2006	2007	2008	2009	2010	2011	2012	2013	2014	2015	2016	2017	2018
Alabama																X
Alaska	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X
Arizona													X	X	X	X
California															X^a	X^b
Colorado		X	X	X	X	X	X	X	X	X	X	X	X	X	X	X
Connecticut													X	X	X	X
Delaware															X	X
District of Columbia															X	X
Georgia		X	X	X	X	X	X	X	X	X	X	X	X	X	X	X
Hawaii													X	X	^c	^c
Illinois														X^d	X^d	X^d
Indiana														X	X	X
Iowa														X	X	X
Kansas													X	X	X	X
Kentucky			X	X	X	X	X	X	X	X	X	X	X	X	X	X
Louisiana																X
Maine													X	X	X	X
Maryland	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X
Massachusetts	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X
Michigan												X	X	X	X	X
Minnesota													X	X	X	X
Missouri																X
Nebraska																X
Nevada															X	X
New Hampshire													X	X	X	X
New Jersey	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X
New Mexico			X	X	X	X	X	X	X	X	X	X	X	X	X	X
New York													X	X	X	X
North Carolina		X	X	X	X	X	X	X	X	X	X	X	X	X	X	X
Ohio									X	X	X	X	X	X	X	X
Oklahoma		X	X	X	X	X	X	X	X	X	X	X	X	X	X	X
Oregon	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X
Pennsylvania														X^d	X^d	X^d
Rhode Island		X	X	X	X	X	X	X	X	X	X	X	X	X	X	X
South Carolina	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X
Utah			X	X	X	X	X	X	X	X	X	X	X	X	X	X
Vermont													X	X	X	X
Virginia	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X
Washington														X^d	X^d	X
West Virginia															X	X
Wisconsin		X	X	X	X	X	X	X	X	X	X	X	X	X	X	X
Total	7	13	16	16	16	16	16	16	17	17	17	18	27	32	37	41

^a Collected data for violent deaths that occurred in 4 counties (n = 1,866; representing 27.8% of violent deaths that occurred in California in 2017), in accordance with requirements under which the state was funded.

^b Collected data for violent deaths that occurred in 21 counties (n = 3,659; representing 55.1% of violent deaths that occurred in California in 2018), in accordance with requirements under which the state was funded.

^c Excluded from data years 2017 and 2018 due to incomplete case reporting.

^d Collected data on >80% of violent deaths in state, in accordance with requirements under which the state was funded.