“Any chronic illness is a curse,” wrote Deborah Danner, a 66-year-old African American woman, in an essay for her adult education class. “Schizophrenia is no different,” she continued, “its only ‘saving grace,’ if you will, is that as far as I know it’s not a fatal disease.” Four years later, in October 2016, Deborah Danner was shot and killed by a police officer who entered her New York City apartment after a call from a concerned neighbor [1]. No saving grace for Ms. Danner, as it turns out.

Encounters between police and people living with mental illness are common. As law enforcement officers have become first responders for mental health emergencies, such encounters have become increasingly significant in our society. Most encounters take place without incident, and police can often prove helpful in navigating moments of crisis. Nevertheless, many communities across the country have experienced high-profile incidents in which officers used force—occasionally deadly force—in encounters with individuals experiencing symptoms of mental illness [2]. Academic psychiatrists, alongside other mental health professionals, have vital roles to play as researchers, leaders, advocates, and teachers in understanding and addressing the very serious issue of officer-involved deaths of people living with mental illness.

Missing and Uncertain Data on

One of the biggest challenges for making informed decisions about how to improve these interactions between the mentally ill and law enforcement officers is the lack of trustworthy and complete data. Advocates estimate that the risk of being killed during a police incident in the USA is “16 times greater for individuals with untreated mental illness than for other civilians approached or stopped by officers” [3]. Rigorously derived source data to substantiate such worrisome estimates are lacking, however. For example, from 2003 to 2014, the Bureau of Justice Statistics was tasked with maintaining a census of “Arrest-Related Deaths,” a category which included homicide by members of law enforcement alongside various forms other causes of death during detention. A 2014 audit of the data collected from 2003 through 2009 revealed that the census captured fewer than half of all homicides by law enforcement, and the data on arrest-related deaths did not meet Bureau of Justice Statistics data quality standards [4]. Activists criticized the census program for failing to mandate the participation of local law enforcement agencies. In March of 2014, the federal data-gathering program was suspended until further notice. Testifying before Congress in 2015, then FBI director James Comey lamented the poor understanding of law enforcement-related deaths, calling the situation “embarrassing,” and stating that “we can’t have an informed discussion because we don’t have data” [5].

Under considerable pressure from lawmakers and the public, the Department of Justice (DOJ) announced a new pilot program in 2016 to collect information about shootings and other violent encounters with the public by federal agents. The DOJ pilot program is limited in scope, but is a sign that the federal government is engaged in activities to ensure public accountability for police-involved shootings. Advocates remain concerned about the future of the pilot program, as well as the priority that current and past federal administrations have given to public accountability of law enforcement [6].

The Washington Post has stepped forward to assemble a database to track police-involved shootings [7]. While this approach is no substitute for the consistent reporting that a coherent, legally mandated federal initiative would entail, the data gathered by the Washington Post is highly revelatory. By aggregating the incidents described in the Washington Post annual reports on police shootings from 2015 to 2017, trends can be identified by population. For example, an overwhelming percentage of deaths from police shootings involve male victims (96%), as shown in Fig. 1. This observation is perhaps unsurprising, given that over 80% of those arrested for suspicion of committing a violent crime are men, according to the FBI [8]. And though the relationship between gender, violent crime, and police shootings is far from simple or transparent, the Washington Post reporting suggests that some relationship does exist and that men have significantly greater risk than women.

Fig. 1
figure 1

Individuals killed by police, per Washington Post database (n = 2945), by race/ethnicity, 2015–2017, compared with base population, US Census Estimates 2016

Full size image

Entries in the Washington Post database also suggest other patterns related to the identity of those killed by police. Looking at the database from 2015 to 2017, for instance, we find that 24% of people shot and killed by police were African American. As African Americans made up 13.3% of the overall population in the 2010 US Census, this suggests that African Americans are overrepresented among those shot and killed by police. In contrast, 48% of individuals shot and killed by police from 2015 to 2017 were white, although non-Hispanic whites make up 61% of the population according to the US census. While more total police shooting deaths involved white people, whites were underrepresented among those shot and killed by police relative to their proportion in the US base population. These patterns (Fig. 2) have stayed remarkably consistent over the three years that the Washington Post has been documenting deaths from police-involved shooting. The large number of deaths in encounters with law enforcement and the differential impact on minority populations lend urgency to the struggle to address the effects of disadvantage and bias based on race and ethnicity and the need for structural changes in policing.

Fig. 2
figure 2

Individuals killed by police, per Washington Post database (n = 2945), by gender, 2015–2017, compared with base population, US Census Estimates 2016

Full size image

Mental Illness as a Risk Factor

Even with the Washington Post data, it is difficult to accurately determine the number of individuals suffering from mental illness who have died in police-involved shootings. For instance, how is “mental illness” recognized and documented in accounts of officer-related deaths? We may have relative confidence in the capacity of law enforcement or local news outlets to accurately report on the ethnicity or gender of an individual killed by the police, but there is little agreement among stakeholders on how to categorize the presence of mental illness in this context. Because mental illness is highly stigmatized, families and neighbors may be reluctant to disclose this information. And because of laws protecting the confidentiality of patients, health professionals will not volunteer such information to police or to members of the media. The Washington Post has made a good faith effort to account for the deaths of individuals who may have had psychiatric conditions by including a category of “signs of mental illness” in their database. The database may not be correct, however, as the Post’s reporters are not trained mental health professionals and must assess for “signs of mental illness” using only police and media accounts.

According to the Washington Post data, about 25% of individuals shot and killed by police from 2015 to 2017 showed signs of mental illness. As about 4% of the total US population lives with serious mental illness and about 18% lives with less severe psychiatric conditions [9], these data suggest that people suffering from mental illness are overrepresented in instances of lethal shootings by police. Moreover, if “signs of mental illness” indicate the presence of serious mental illness rather than less severe conditions, individuals with serious mental illness are highly overrepresented.

Individuals with mental illness who belong to other underrepresented minority groups appear to be at greater risk of being killed in an encounter with law enforcement. According to the Washington Post database, 2945 people were fatally shot by police from 2015 to 2017. One would expect, based on the US population, that 11 to 78 of these people would be African American and have a mental illness. The Post identified 113 individuals shot and killed by police from 2015 to 2017, however, who both were African American and had “signs of mental illness.” If showing “signs of mental illness” is assumed to be equivalent with having a serious mental illness, the number of African American individuals suffering from serious mental illness who were killed by police outpaced the proportion of that group in the general population by a factor of 10:1, a ratio that is as astonishing as it is devastating.

Comparing the proportion of individuals killed by police from 2015 to 2017 who show signs of mental illness versus those who do not show signs of mental illness by race/ethnicity reveals a somewhat different narrative (see Fig. 3). African Americans and people suffering from mental illness are both overrepresented in the number of people killed by police. However, African Americans showing signs of mental illness make up a relatively small proportion of the total number of African American people killed by police when compared with the proportion of white people showing signs of mental illness who were killed by police. In other words, suffering from mental illness does not place an African American individual at significantly greater risk. By contrast, while non-Hispanic white people make up a relatively small percentage of the total population shot and killed by police overall, white people who showed signs of mental illness make up a large percentage of total non-Hispanic white deaths (32%) in police encounters.

Fig. 3
figure 3

Proportion of individuals killed by police, per Washington Post database (n = 2945), who show signs of mental illness vs. those who do not show signs of mental illness, by race/ethnicity, 2015–2017

Full size image

A similar dynamic emerges when we compare deaths of those who showed signs of mental illness across gender (see Fig. 4). While relatively few women compared to men are killed by police, the Washington Post data reveals that a very large percentage of women killed by police from 2015 to 2017 showed signs of mental illness (37%). Multiple factors and reporting discrepancies may be at work in these numbers and more complete data and explanatory models are sorely needed. Even so, the trend line appears to indicate that although white people and women are generally less at risk for police violence, this experience does not extend to white people and women who suffer from mental illness.

Fig. 4
figure 4

Proportion of individuals killed by police, per Washington Post database (n = 2945), who show signs of mental illness vs. those who do not show signs of mental illness, by gender, 2015–2017

Full size image

Attempts to visualize the disproportionate burden of police violence borne by multiply vulnerable communities must be accompanied by a number of caveats. First, there are numerous difficulties in attempting to derive a clinical evaluation of a deceased individual based on scant empirical evidence or narrative accounts. Additionally, a number of mental health professionals have expressed concerns about the tendency to use a label of “mental illness” as an explanatory device to account for violent crimes which are particularly horrific, such as mass shootings, despite the fact that mental illness contributes very little to overall violence in our society [10, 11]. Moreover, attempts to track outcomes of law enforcement encounters with individuals suffering from mental illness could easily introduce new and serious negative effects, such as further stigmatization. In light of rhetoric which seeks to distract from the root causes of violence by scapegoating mentally ill populations in the aggregate, it is particularly important that we take care to avoid attribution of serious mental illness to individuals for whom we have no way of making clinical evaluations. And still, as an overall picture, it is clear that as a society we are facing significant challenges in relation to how law enforcement officials interact with individuals who are potentially vulnerable to adverse outcomes due to mental illness.

These complexities, and the broader need for reliable data, highlight the need for the field of academic psychiatry to come forward to convene and sustain a substantive conversation among stakeholders about mental illness and encounters with law enforcement. Psychiatric leaders and researchers may play a valuable role by bringing together population scientists along with patient advocates, law enforcement partners, and members of groups that are directly affected to explore how best to develop systems for data collection, sharing, and interpretation. All of these facts highlight the critical need for better data collection infrastructure and classification approaches that are shared among stakeholders.

It is worth examining the frequency with which fatal incidents involving police befall individuals whose mental illness co-occurs with another source of potential vulnerability. Many victims of police violence who suffer from mental illness, for instance, are also veterans, are marginally housed, are gender or sexual minorities, or come from low-income communities. Such co-occurrence is likely indicative of our mental health system’s inability to adequately care for diverse segments of our population. Understanding the complex relationships between systemic failures to protect vulnerable populations and tragic outcomes for specific individuals warrants substantial and sustained inquiry by academic researchers and policy makers, as well as by the news media.

Furthermore, an intersectional approach to the evidence reveals that race/ethnicity and mental illness function as compounding risks for fatal outcomes with law enforcement. This observation brings up troubling questions regarding correlations between the disproportionate impacts of police encounters on people of color and lower rates of access to mental health care among people of color. Could the deaths of people of color by police be less the result of biased attitudes of individual police officers and more indicative of inequities in our mental health delivery systems?

When we set aside the statistics and begin to research narrative accounts of instances in which people suffering from mental illness died in encounters with law enforcement officers, overlapping vulnerabilities reveal themselves. Examples of media accounts of police-involved deaths of people living with mental illness are provided in the Teaching Tool appended to this editorial. This Appendix Table 1 provides insight into the experiences of those who lost their lives and may be a valuable educational resource in residency curricula to help shape the foundational knowledge of early career psychiatrists.

While there is an urgent need for better data and more robust analyses in the efforts to inform and reform public practice and public policy, the statistics and newspaper stories cannot convey the nuances of the individual experiences of vulnerability and fear. For these reasons, it is our hope that these very compelling aspects of the experience of encountering law enforcement by people with mental illness will be considered and discussed by residents and fellows whose careers will be dedicated to the care of this special population.

Positive Steps, Looking Ahead: Roles for Psychiatrists

Deborah Danner should be alive right now, period.

If the protocols had been followed, she would be alive. It’s simple as that.

—New York City Mayor Bill DeBlasio [12]

Psychiatrists can help address and advance understanding of the devastating phenomenon of mentally ill individuals being disproportionately represented among police-involved deaths in our society. Academic psychiatrists may undertake research and collaborate with others to develop data collection systems that are trustworthy and can illuminate causes and contributors of death for people living with mental illness. Psychiatrists in academic and community settings can engage deeply and substantively with their neighbors and government representatives, fostering partnerships that will address the most salient issues in their stakeholder communities.

Psychiatrists, alongside other mental health professionals, can develop evidence-based strategies and educational activities and programs to help law enforcement officers who have been forced to assume the role of first responders in very difficult situations with individuals living with mental illness. Psychiatrists can also work within their systems of care to improve access for mentally ill individuals at greatest risk and who have multiple sources of potential vulnerability (e.g., veteran, homeless or marginally housed, and underrepresented minority individuals). Very importantly, academic psychiatrists can seek to inform public policy and can teach their colleagues and their students about the issues faced by mentally ill individuals in encounters with police.

Being of service to help support law enforcement officers who themselves may become traumatized by their experiences and are often placed in nearly impossible positions is yet another important role for psychiatric practitioners. Indeed, there have been considerable efforts, often principally led by law enforcement officers who have been thrust into the role of first responders in psychiatric emergency situations, to address these tremendous challenges in encounters between law enforcement and people living with mental illnesses. Programs such as Crisis Intervention Trainings [13] may have beneficial effects on relationships between law enforcement and mentally ill citizens. Psychiatrists in diverse settings are important partners, as teachers and as co-learners, in these valuable efforts.

Looking ahead, there are many opportunities to positively influence the broad and significant societal problem of premature mortality among people with mental illness. Mental disorders are the second-leading cause of premature mortality and disability in the world [14,15,16]. People with mental illnesses are at greater risk for developing co-occurring physical health conditions, for being unsheltered, for poverty, for being victims of violence, and for being victims of police-related shootings. Such problems are under-recognized and represent, in our view, the ultimate health disparity for people living, and dying, with mental disorders. It is up to us to ensure that this grave inequity does not continue unnoticed so that, someday, mental illness will have the “saving grace” that Ms. Danner believed.