Search tips
Search criteria 


Logo of jurbhealthspringer.comThis journalToc AlertsSubmit OnlineOpen ChoiceThis journal
J Urban Health. 2016 April; 93(Suppl 1): 8–31.
Published online 2015 December 17. doi:  10.1007/s11524-015-0008-7
PMCID: PMC4824690

Who Shot Ya? How Emergency Departments Can Collect Reliable Police Shooting Data


This paper examines an alternative solution for collecting reliable police shooting data. One alternative is the collection of police shooting data from hospital trauma units, specifically hospital-based violence intervention programs. These programs are situated in Level I trauma units in many major cities in USA. While the intent of these programs is to reduce the risk factors associated with trauma recidivism among victims of violent injury, they also collect reliable data on the number of individuals treated for gunshot wounds. While most trauma units do a great job collecting data on mode of injury, many do not collect data on the circumstances surrounding the injury, particularly police-involved shootings. Research protocol on firearm-related injury conducted in emergency departments typically does not allow researchers to interview victims of violent injury who are under arrest. Most victims of nonfatal police-involved shootings are under arrest at the time they are treated by the ED for their injury. Research protocol on victims of violent injury often excludes individuals under arrest; they fall under the exclusion criteria when recruiting potential participants for research on violence. Researchers working in hospital emergency departments are prohibited from recruited individuals under arrests. The trauma staff, particularly ED physicians and nurses, are in a strategic position to collect this kind of data. Thus, this paper examines how trauma units can serve as an alternative in the reliable collection of police shooting data.

Keywords: Police, Brutality, Hospital, Emergency department, Violence, Intervention

I’ve never had someone say they were shot by the police who wasn’t and I’ve been doing this for 25 years.

Dr. Roy Smith (Trauma Surgeon)

I’ve heard from a number of people who have called on policymakers to ensure better record keeping on injuries and deaths that occur at the hands of the police.

Former US Attorney General Eric Holder

Demographic data regarding officer-involved shootings is not consistently reported to us through our Uniform Crime Reporting Program. Because reporting is voluntary, our data is incomplete and therefore in the aggregate, unreliable. The first step to understanding what is really going on in our communities is to gather more and better data related to those we arrest, those we confront for breaking the law and jeopardizing public safety, and those who confront us. Data seems a dry and boring word but, without it, we cannot understand our world and make it better.

James Comey, Director of the FBI

Injury and death from excessive use of force by police officers is classified as a human rights violation by the World Health Organization.1,2 Despite the classification of excessive force as a form of violence, medical researchers have produced scant data characterizing law enforcement-perpetrated abuse.3 This paper examines the current state of data collection on fatal and nonfatal police officer-involved shootings. Using in-depth interviews collected from trauma staff at Prince George’s Hospital Trauma Center, the busiest Level II trauma center in USA, we examine the perceptions of trauma staff regarding the collection of police-involved shooting data by hospital violence intervention programs (HVIPs). We conclude with programmatic and policy recommendations regarding how HVIPs can collect accurate data on police-involved shootings.

In a January 2015 speech at the Justice Department honoring Martin Luther King, Jr., Attorney General Eric Holder stated “the troubling reality is that we lack the ability right now to comprehensively track the number of incidents of either uses of force directed at police officers or uses of force by police, obtaining better data on police shootings as well as police officer deaths, would represent a common-sense step to address serious concerns about the need to safeguard civil liberties…this [lack of data] strikes many including me as unacceptable.”

Six months prior to Attorney General Holder’s remarks on MLK day, in the summer of 2014 following the protests in Ferguson, MO, surrounding the fatal police-involved shooting of Michael Brown, an unarmed 18-year-old African-American male, Director of the Federal Bureau of Investigations, James Comey, asked his staff for figures on the number of Blacks who had been shot by the police. His staff replied that there were no uniform statistics on such shootings due to the lack of collecting and sharing of better information about violent encounters between the police and citizens. The recent spate of highly publicized fatal police shootings of Black youth and young Black men in USA has ignited a national conversation on police brutality. One issue that has gained considerable attention is the lack of reliable data recorded and reported by police departments on police-involved firearm-related homicides and shootings. To date, there is no reliable data on how many people are fatally and nonfatally shot by the police every year. In a recent article published in the Washington Post4 which criticized the lack of reliable data on police shootings reported to the FBI by police departments in USA, the article reported the following:

Criminal justice experts note that, while the federal government and national research groups keep scads of data and statistics— on topics ranging from how many people were victims of unprovoked shark attacks (53 in 2013) to the number of hogs and pigs living on farms in the U.S. (upwards of 64,000,000 according to 2010 numbers) — there is no reliable national data on how many people are shot by police officers each year. How many people in the United States were shot, or killed, by law enforcement officers during that year? No one knows.

Data gathered by the Federal Bureau of Investigation (FBI) on police shootings indicates there are approximately 400 justifiable police-involved homicides each year; however, independent researchers have found that the number of police-involved fatal shootings of citizens is closer to 1000 homicides per year.4 Police-involved homicides reported by independent researchers are roughly 2.5 times higher than the number of homicides reported by the FBI. Furthermore, there is no reliable or accurate data on police-involved nonfatal shootings. Criminologists who conduct research on police-involved shootings have found that the data on nonfatal police-involved shootings is unknown and is considered by many researchers as a “black hole.”4

The Department of Justice keeps no comprehensive database or record of police shootings from the roughly 18,000 law enforcement departments in USA. Furthermore, there is no federal mandate that law enforcement departments report police-involved shooting data to the FBI. Under current federal laws, there is nothing requiring any of the 18,000 police departments and other law enforcement agencies across the country to report to the public or the Justice Department anything about shootings involving officers.5 Law enforcement departments are only required to “self-report” shootings. To further emphasize this point, in 2012, only 750 (4 %) of the nation’s 18,000 law enforcement agencies reported police-involved shooting data to the FBI.6

Although approximately 91 % of law enforcement departments and agencies in USA voluntarily report crimes such as homicide, rape, robbery, assault, theft, and burglary to the FBI (documented in the Uniform Crime Reports (UCR)), there is no category in the UCR which captures an officer’s use of force that was deemed not legally justified. Furthermore, there is no category to report police-involved shootings where the person shot by a police officer was not killed. However, there is a category classified as “justifiable homicide by weapon by law enforcement officer,” which records when a felon is killed by a law enforcement officer in the line of duty. The FBI also collects data concerning the felonious and accidental line of duty deaths and assaults of local, state, tribal, and federal law enforcement officers. According to the UCR, the data on law enforcement officers killed and assaulted includes weapons used, the presence of body armor, and the circumstances surrounding murders of and assaults on officers. In addition, the data concerning officers killed in the line of duty includes detailed information about the victim (e.g., age, sex, race, height, and number of years of law enforcement) and information about the assailants (e.g., age, sex, race, criminal histories, and the number of assailants arrested). In addition, there are narratives that provide detailed summaries concerning the death of each officer feloniously killed in the line of duty that year.

Other organizations such as the National Law Enforcement Memorial Fund (NLEOMF) maintain research departments that serve as a national clearinghouse of information and statistics on law enforcement line of duty deaths. NLEOMF publishes an annual law enforcement fatalities report. This report documents the circumstances of fatal shootings (e.g., ambush, traffic stop or pursuit, investigating suspicious circumstances, disturbance calls, attempting arrest, investigative activity, accidently shot, burglary, drug-related matters, robbery, and tactical). The report also includes both felonious and nonfelonious fatalities. Thus, the data collected on law enforcement officers killed and assaulted in the line of duty are quite comprehensive. Despite the comprehensiveness of the data collected by the NLEOMF on officers killed in the line of duty, the Fraternal Order of Police (FOP), the largest national professional police organization in USA, does not collect data on officers killed in the line of duty nor does it collect data on officer-involved shootings. An exhaustive review of the FOP research website found that much of the data the FOP collects focuses on labor-related issues. Thus, for civilians killed or injured by law enforcement, this data is scant.

Because there is no federal mandate requiring law enforcement to report police-involved shootings to the Justice Department, this process has significantly resulted in the FBI underreporting the number of police-involved homicides. For example, the independent research organization which maintains a database of all deaths through police interaction documented 4813 arrest-related deaths by law enforcement (2003–2009). However, the Justice Department recorded 2657 justifiable homicides by the police during the same period. Other independent studies have also found discrepancies between the number of police-involved killings recorded by law enforcement and those recorded by independent organizations. A Wall Street Journal analysis of police killings conducted in December 2014 from 105 of the country’s largest police agencies found more than 550 police killings during those years that were missing from the national data.7 The Wall Street Journal (WSJ) requested internal records on killings from the nation’s 110 largest police departments; 105 of those departments provided data.

The data collected by the WSJ indicated that at least 1800 police killings in those 105 departments between 2007 and 2012 were 45 % higher than the FBI’s total for justifiable homicides by those departments’ jurisdictions. Furthermore, justifiable police homicides from 35 of the 105 agencies contacted by the WSJ did not appear in the FBI records at all. Some of the agencies such as the Fairfax County Police Department in Virginia (a department with approximately 1370 sworn officers) stated that they did not view justifiable homicides by law enforcement officers as reportable data. Some of the discrepancies in underreporting are the result of large jurisdictions that do not routinely report police-involved deaths of civilians. The Washington DC Police Department did not report any police-involved homicides to the FBI for an entire decade beginning in 1998 the same year that the Washington Post found that the city had one of the highest rates of officer-involved killings in the country. The New York City Police Department (the largest in the nation) reported no police-involved homicides to the FBI from 2007 to 2012; however, the WSJ collected data from the New York City Police Department (NYPD) which reported 68 police-involved homicides during this time period.

According to an analysis of police-reported shootings conducted by ProPublica,8 law enforcement departments in the state of Florida have not filed reports since 1997. Reports from the Florida Department of Law Enforcement do not conform to FBI requirements and have not been included in the national data since 1996. For example, the Miami-Dade Police Department reported no police-involved homicides to the FBI from 2007 to 2012, yet the data collected by the WSJ reported 53 homicides during this period. Some states such as Florida and New York attribute their inability to report data to the FBI to outdated technology and software.

Furthermore, the data collected on police-involved shootings by other governmental agencies varies widely from the data collected by the FBI and other independent organizations. The Centers for Disease Control and Prevention (CDC) collects data on police-involved homicides using death certificates. The CDC data reported by coroners and medical examiners records the medical cause of death, not the legal justification of it. Death certificates may or may not include details about officer involvement. The CDC does not evaluate whether police-involved shootings were justifiable. This evaluative approach is deferred to the FBI. To highlight the discrepancies in the reported totals, we analyzed police-involved fatal shooting data taken from the year 2012 collected by the FBI and CDC. According to the UCR in 2012, there were 409 police-involved fatal shootings compared to 196 police-involved fatal shootings recorded by the CDC for the same year.

While the data collected by the FBI, Bureau of Justice Statistics (BJS), and the CDC varies widely on fatal police-involved shootings, there is literally no data on nonfatal police-involved shootings. This data is critically important because there are far more nonfatal shootings in USA than firearm-related homicides. For example, the most recent police firearm discharge reported by the NYPD (nation’s largest police department) in 2013 found 40 incidents where officers intentionally discharged their firearms during adversarial conflict.9 In total, eight citizens were shot and killed by officers and 17 were shot and injured. However, there is no category used by the FBI, BJS, or the CDC to capture how many people are nonfatally shot by the police every year. A recent investigation by the Justice Department on police shootings by the Philadelphia Police Department (nation’s fourth largest police department) found that 400 citizens were shot by the Philadelphia police between 2007 and 2013. On average, this is approximately one civilian shot each week for a period of 7 years. Ironically, during this period in Philadelphia when violent crime was dropping, police shootings were increasing. While 400 police-involved shootings may seem high for a 7-year period, it is impossible to put these 400 shootings into context whether these numbers are high or low because there is no reliable national data on police officer-involved shootings. However, recent federal legislation may address the lack of reliable reporting of police shooting data.

Death in Custody Reporting Act

In December 2014, Congress passed the Death in Custody Reporting Act. The Deaths in Custody Reporting Program (DCRP) collects data on deaths that occur in the process of arrest or while inmates are in the custody of local jails or state prisons. Arrest-related mortality data are collected through a centralized state reporting coordinator. Local jail and state prison data are collected directly from jails and state departments of corrections. The DCRP provides individual-level data on the number of deaths by year, cause of death, and decedent age, race or Hispanic origin, and sex. These data are also used to produce facility and population mortality rates. The DCRP began in 2000 under the Death in Custody Reporting Act of 2000 (P.L. 106-297), which required the collection of individual data on deaths in the process of arrest, local jails, and state prisons. While this legislation expired in 2006, BJS has continued to collect these data, as they represent a unique national resource for understanding mortality in the criminal justice system. This bill also requires any law enforcement agency receiving federal funding to record and report all police-involved shootings to the Department of Justice. Agencies that do not report police-involved shootings may have their federal funding revoked. Criminal justice experts contend that it will take years to perfect the collection of data. Although the DCRP collects data that occur in the process of arrest, police-related shootings may occur when an individual is not being placed under arrest. For example, several high-profile police-related shootings in 2015 occurred during traffic stops. The shooting death of Samuel Dubose by a University of Cincinnati police officer occurred during a routine traffic stop, similarly the shooting of Walter Scott in North Charleston, South Carolina, where the victim was fired upon eight times after fleeing from Officer Michael Slager during a routine traffic stop. While the DCRP collects data on deaths, it does not collect data on nonfatal police-involved shootings. The DCRP cannot account for how many people are nonfatally shot by the police every year. Furthermore, the Death in Custody Reporting Act has been largely ineffectual because the threat of withholding federal funding from states for not reporting data has not resulted in compliance. One aspect of the Death in Custody Reporting Act (DRCA) that has been problematic is that the act does not provide states with funding to design or collect data. Because the majority of police departments (80 %) have less than 10 officers, departments usually do not have the capacity to collect and manage data. As a result, the DCRA is basically an unfunded mandate.10

This paper specifically focuses on the lack of data collected by law enforcement on citizens who survive officer-involved shootings. The DCRP neglects to account for those individuals who survive police officer-related shootings. This is critically important because the majority of citizens fired upon by the police survive. In a recent article published by the Washington Post on police-involved shootings, criminologist David Klinger at the University of Missouri at St. Louis, an expert on police-involved shootings, found that “studies show that most of the people who are shot by the cops survive and most of the time when cops shoot, the bullets don’t hit.” According to Klinger, a recent study on police-involved shootings in St. Louis found that between the years 2003 and 2012, there were 230 instances when officers fired their weapons and only 37 of those fired upon were killed. Klinger notes “if your statistics just look at dead bodies you’d be under-counting police involved shootings by 85 percent.” Presently, the DRCP does not capture data on nonfatal shootings. What other ways can we collect accurate and reliable data on police-involved homicides and nonfatal shootings?

Emergency Medicine and the Police: Why Accurate Data Collection Is So Important

Despite growing recognition of violence and its health consequences and the World Health Organization’s classification of police officers’ excessive use of force as a form of violence, public health investigators have produced scant research characterizing police-perpetrated abuse.3 Every day in USA, law enforcement officers and emergency physicians cross paths when a victim or offender presents to the emergency department for care.11 However, most assaults, including serious violent injury (e.g., firearm-related assaults) reported and recorded by the police, do not match assault data recorded by hospital emergency departments (EDs).12 Studies have reported that EDs detect far more assaults in the community than those recorded by the police.12 In studies conducted in the UK, it has been found that only about 25–50 % of assaults that result in medical treatment are recorded by the police and that recording cannot be predicted on the basis of the severity of injury.13 In one study, which compared ED data versus police identification of an assault, the ED recorded twice the number of victims of assault as the police.14 Studies conducted in the UK and Scandinavia that matched data from EDs and police have shown that only a quarter to one third of violent incidents that resulted in treatment by an ED appear in police records.15

In USA, 13 % of shootings resulting in ED care in Atlanta were not included in citywide police records.16 Researchers found that a number of medically documented nonfatal shootings could not be matched with a corresponding police report. Several cases involved severe or multiple gunshot wounds. A three-city study found that 9 % of medically documented cases of gunshot injuries could not be matched with a corresponding police report.16,17 Other US studies on the recording of violence by medical and police services suggest that a large majority of assaults resulted only in ED recording and not by the police.13 For example, low rates of police recording intimate partner violence have prompted legislation in some states to make reporting by health professionals’ mandatory.18

Although most states require medical personnel to report all nonfatal shootings to law enforcement, the rate of compliance is unknown.16 For example, the state of Georgia requires health-care workers to report suspected cases of firearm assault to local law enforcement agencies but it does require reporting to a regional or state agency. Consequently, health-care providers frequently fail to report shootings.16

While the reporting of violent assaults that result in treatment by the ED is low among law enforcement, reporting by medical staff of some violent offenses that result in treatment is also low. For various reasons, some health-care providers choose not to report. For example, busy Level I and Level II trauma centers that experience high rates of patient volume may assume that another member of the trauma staff is recording and reporting the injury. This often results in missing reports. In other instances when reports are created, the data may be inaccurate. Reports may have the medical record number but no name. This results in reports with missing data. Some EDs may not have the technological software to link medical and police reports so they choose not to report at all.16

An exhaustive review of studies on the recording of violence by medical and police services found no studies which captured how nonfatal police-involved shootings were recorded. Furthermore, there appears to be almost no international, systematic, ethically sound ED and police collaboration.13 For firearm injury surveillance in USA, although systems are being developed or enhanced (e.g., Georgia Firearm Injury Notification), the systematic study of the extent to which there is correspondence between ED and police recording has not happened.

How ED physicians perceive and record cases of police abuse is also problematic. Most ED physicians have not received didactic training on the management of cases of suspected excessive use of force, and very few residencies provide any type of instruction.1 One large study on emergency physicians’ (EP) perceptions of excessive force used by the police found that most EPs believe they have seen cases of excessive force but failed to report them. However, over 70 % felt that excessive use of force by the police should be reportable events.1 In another study, half of emergency physicians, nurses, and clinical and managerial staff believed that they should have a role in victim protection and support, the detection of crime, and community crime prevention.19 Yet, due to the low and disparate rates of reporting by law enforcement and physicians, little is known about the epidemiology of nonfatal police-involved shootings. Injury-derived data from EDs are the only continuous source of information about victimization in violent crime and should be used to supplement police data.

Police Brutality, Race, and Gender

Police shootings also disproportionately impact specific demographic groups, particularly young Black males. The police-involved shooting deaths of unarmed young Black men over the past two decades have highlighted the disturbing pattern of police brutality among this population. The high-profile police-involved shooting deaths of Amadou Diallo (Bronx, NY), Sean Bell (Queens, NY), Oscar Grant (Oakland, CA), Michael Brown (Ferguson, MO), Tamir Rice (Cleveland, OH), and Walter Scott (North Charleston, SC) represent some of the fatal police-involved shootings that have garnered national attention. Young Black men are at far greater risk of being shot by the police than their white counterparts. Data collected by the FBI on police shootings from 2010 to 2012 indicates that young Black males were 21 times more likely to be shot and killed by the police than young White males.8 The 1217 fatal police shootings from 2010 to 2012 reported in the federal data show that Blacks aged 15–19 were killed at a rate of 31.17 per million while 1.47 per million white males in the same age range died at the hands of the police. Between the years 1980 and 2012, 41 teens 14 years or younger were reportedly killed by the police: 27 (66 %) were Black, 8 were white (20 %), 4 were Latino (10 %), and 1 was Asian (2 %). Of the 15 teens shot fleeing arrest from 2010 to 2012, 14 were Black. Qualitative studies on recurrent trauma among young Black men found that harassment and racial profiling by the police were significant sources of trauma and stress.20 Researchers found that young Black men expressed a lack of faith in the police and that they could not depend on the police to protect them from danger.

Hospital Violence Intervention Programs

There is potential for EDs to contribute substantially to documenting excessive police force and to implement innovative approaches to stem unchecked forms of police brutality.12 EDs have the opportunity to advocate on behalf of those injured by the police. One alternative ED resource for collecting reliable police shooting data which has largely been ignored in the discourse on police shooting data are HVIPs. These programs situated in Level I and Level II trauma units in many major metropolitan areas throughout USA (e.g., Los Angeles, Chicago, Philadelphia, Boston, and Baltimore) are tasked to reduce violent injury among patients presenting to the ED for gunshot wounds, stabbings, or assaults. Presently, there are 36 HVIPs that are members of the National Network of Hospital-Based Violence Intervention Programs; these also include three programs, in Canada (two) and in London, England (one). Network members include trauma surgeons, program directors, researchers, medical directors, trauma staff, social workers, and behavioral health clinicians who collect data and also share their experiences and perspectives to create effective strategies to reduce violence and violent injury.

HVIPs are dedicated to engaging patients during the “window of opportunity” when they are recovering in the hospital after a violent injury. The window of opportunity is used to reduce the likelihood of retaliation and recurrence of violent injury. While the primary goal of these programs is to reduce the risk factors and costs associated with violent trauma recidivism, they also collect accurate and reliable data on the number of individuals treated for intentional firearm-related injuries. Several studies have found that HVIPs are effective in collecting violent injury data and that they can reduce repeat violent injury and the costs associated with injury.2123

These programs collect data on the mode of injury (i.e., gunshot wound or GSW) as well as data on the context and circumstances surrounding the violent injury. HVIPs collect data on the circumstances of the injury in ways similar to data collected by the FBI and NLEOMF on officers’ shot in the line of duty. The data were collected, but HVIPs can provide an opportunity to understand the context of police-involved shootings. As co-directors of an emerging HVIP, we believe that HVIP trauma staff and researchers are in a strategic position to collect accurate and reliable data on police shootings. Given the low rates of police recording offenses that result in ED treatment, HVIPs are positioned to collect reliable data on police-involved shootings. Furthermore, numerous studies indicate that local law enforcement agencies are often reluctant to self-report police-involved shootings to state and federal authorities. We argue that at least three sources of data collection are needed to record and report both fatal and nonfatal police-involved shootings: police, ED (includes emergency medical services (EMS), and medical examiner reports. This paper examines how HVIPs can serve as a primary resource in the reliable collection of police shooting data.


The data used for this paper were collected as part of a larger clinical qualitative research study on the risk factors for recurrent violent injury, linkages, and barriers to care and HIV risk behaviors among 25 [N = 25] young Black men (aged 18–34) treated by the Prince George’s Hospital Trauma Center (PGHTC) for violent injury (e.g., gunshot, stabbing, or assault). PGHTC is the busiest Level II trauma unit in USA and treats on average 700 victims of violent injury per year. The trauma unit treats approximately 200 firearm victims per year (29 % of violently injured patients). Twenty-one percent of all trauma patients treated by PGHTC are treated for violent injury. This ratio is highest in the state of Maryland. The larger qualitative study explored the risk factors that contribute to repeat violent injury among young Black men. Understanding the context of repeat violent injury is important because the rate of repeat violent injury ranges from 35 to 44 % with a 20 % 5-year mortality rate. In order to develop strategies to prevent repeat violent injuries among young Black men, a population disproportionately over-represented among victims of violent injury, it is important to understand the risk factors for recurrent violent injury. However, little is known qualitatively about these characteristics among young Black male victims of firearm-related injuries.20,24 This longitudinal qualitative analysis was undertaken to investigate significant individual and environmental risk factors associated with repeat ED visits for violence among young Black men. In addition, the study aimed to investigate how young Black men accessed continued health care for their traumatic injuries (both physical and psychological) following discharge from the hospital. The final question the study aimed to address is whether young Black male victims of violent injury engage in HIV risk behaviors. This question was specifically germane for the context of this study because Prince George’s County has the highest rates of new HIV infections in the state of Maryland. Young Black men are the majority of new HIV infection cases in the county. Understanding and addressing these factors will enable practitioners and health systems to develop targeted strategies to prevent repeat visits for intentional violent injuries, develop critical linkages to continued care, and provide HIV testing for victims of violent injury.25

The research team is composed of three African-American males with extensive experience conducting research on violence and trauma among low-income young Black men. One member of the research team is the Chief Medical Officer of PGHTC. He is also a trauma surgeon and founding director of the University of Maryland Medical Systems Violence Intervention Program, one of the nation’s first HVIPs. Two members of the research team are ethnographers with extensive experience conducting research on the social context of violence among low-income Black male youth and young adults.

The ethnographers spent a significant amount of time in hospitals’ trauma units recruiting violently injured young Black men at bedside and working closely with trauma staff as they provided treatment and care for our target population. For our study, the hospital IRB required that the research team could not recruit violently injured patients who were under arrest. During the recruitment phase of the study, there were several GSW victims who we were unable to recruit or speak to about the study because they were under arrest. Police officers guarded their rooms 24 hours a day. Only trauma staff personnel were authorized access to their rooms. This raised questions among members of the research team regarding our lack of access and our inability to collect data on the circumstances of their injuries. How many patients were we prevented from speaking to because there were legal ramifications associated with their injuries (e.g., police-involved shooting)? To answer these questions, we chose to interview the trauma staff, specifically staff affiliated with the hospital trauma unit’s emerging HVIP about their perceptions and challenges regarding the collection of police shooting data from victims of violent injury. Furthermore, we chose to interview the trauma staff because they had unfettered access to this population.

The data for this paper was collected from semi-structured interviews and informal conversations with five PGHTC staff: (2) emergency medicine physicians (one physician was the medical director of trauma services); (2) trauma nurses (one nurse was assigned to the violence prevention and trauma outreach program for the hospital); and (1) co-director of an emerging HVIP at PGHTC. Each participant reported on their perceptions of the challenges, barriers, and strategies for collecting police-involved shooting data via the ED. We elected to limit our data to the responses given by the trauma staff because health-care practitioners and health system administrators, specifically health-care providers who work with vulnerable populations of violently injured young Black men, have the greatest potential to develop data collection strategies, systems, and protocols to determine if patients were involved in police-related shootings. While we acknowledge that the responses of patients are important, our analysis focuses specifically on the responses of caregivers who treat violently injured populations in a clinical setting. Our selection of the caregivers was framed by those individuals the research team observed who worked closely with victims of violent injury. In addition, all of the caregivers were affiliated with the emerging HVIP.

We utilized a grounded theory approach.26 Using the grounded theory approach, we generated theory from the data. In contrast to testing a theory or hypothesis, the grounded theory allows the researchers to develop a set of questions or areas of observations to examine a specific social phenomenon. The method is one of discovery. Unlike other methodological approaches where the analysis begins following the completion of the study, using grounded theory, the analysis is initiated as soon as data is collected from the first interview or observation. The analysis produces concepts grounded in the data. Every concept brought into the study or discovered in the research process is at first considered provisional. Concepts become basic units of analysis. Concepts are constantly compared. A concept earns its way into theory by being repeatedly present in each interview or observation (thematic). Concepts are grouped into categories and sub-categories. These categories serve as the foundation for generating and developing theory.

Transcripts and field notes were coded and analyzed using qualitative software Dedoose. Throughout the coding process, the authors discussed and deliberated over emerging codes, categories, and their relationships until a consensus was reached. Using Shephard’s12 framework for emergency medicine and police collaboration to prevent community violence, the following themes were identified: attitudinal, logistical, ethical, and legal.



Some trauma staff expressed a reluctance and unwillingness for hospitals to cooperate with collecting detailed data on police-involved shootings. Here, Dr. Jones, a trauma physician, elaborates on why many hospitals may choose not to collect data:

Hospitals are not going to be willing to cooperate as a general rule.

When asked with more probing questions regarding why Dr. Jones thought hospitals would be unwilling to cooperate, Dr. Jones stated:

Hospitals will not want to do what is outside of the realm of what their mission is. It (collecting police involved shooting data) does not add to the care of the patient at all. It is not data we need.

In a separate interview with Dr. Smith, Medical Director of Trauma Services, he reiterated that some hospitals may be reluctant to collect this kind of data:

Who did it (violent assault) won’t make a difference to the care of the patient nor is it timely. We never ask the question who shot you? Because it is not important to their (patients) care. I don’t ask that question because if you are shot by the police or someone who sold drugs, I have to save your life. It doesn’t matter to me where the bullet comes from. Who did it? While that is an important question, we still have a lot of people dead. Who shot you is one of things we won’t ask because it will not make a difference to the patient’s care at that time. However, it certainly may make a difference after… if they survive.

Kim, a violence prevention and trauma outreach nurse at PGHTC, echoed similar sentiments:

Trauma doesn’t care about the context of what happened because they (staff) will not be able to get a reliable account. A police involved shooting could be an accident or criminal.

Daphne, a trauma nurse with extensive experience working as a violence and trauma outreach nurse at Level I trauma centers in the local metro area, also discussed the reluctance of the trauma staff, particularly nurses, to inquire about the circumstances of the injury:

Often times especially for domestic disputes the context of the injury should be documented, while I think police involved shooting data is necessary, there may be some push back because the hospital may feel there are legal ramifications. Some nurses may be concerned about being a witness in court so they will be hesitant to document or identify it.

Regarding hospitals collaborating with the police to share and exchange data, Dr. Jones reported that police-involved shootings should be investigated, recorded, and reported by the police. Dr. Jones perceived that hospital’s involvement in collecting police-involved shooting data fell outside of the scope of the hospital’s mission:

Hospitals will say all investigations should be done the police because investigating police involved shootings takes hospitals out of what they see as their mission.

Carlton, co-director of the emerging HVIP at PGHTC (and a licensed social worker), had a much different perspective on the hospital’s role in documenting police shootings:

When the police bring someone to the hospital that needs to be treated, the hospital should be able to record and report not only the mechanism of injury, they should also ask and document whether the shooting was the result of someone else or was it by the police? The hospital serves as a good source for a counter-narrative. Right now the police, the UCR and the CDC are not giving the right numbers so let’s provide another data source.

Dr. Smith discussed how the majority of patients he treats for gunshot wounds are not shot by the police but by other young Black men. In his estimation, because he sees so few patients injured in police-involved shootings, the hospital should target its efforts towards reducing the number of young Black men shot by other young Black men:

The majority of people are shot by a person of their own race, so Black men are typically shot by Black men. We can focus on the 10 percent who are shot by the police or the 90 percent shot by another Black man. I want to focus on the 90 percent. If we miss the mark by focusing on the 10 percent we ignore the larger problem. The vast majority of people that come in shot are shot in their own community, it is not police related, so we have not been forced to confront this issue. I agree that people designated to protect you shouldn’t be killing you, but that is not the biggest problem in the Black community.


Participant responses varied considerably on whether it was feasible to collect police-involved shooting data. PGHTC collects this kind of data but not in a systematic or standardized approach. Here, Dr. Jones describes one approach to collecting police-involved shooting data. According to Dr. Jones, this data may be found in the patient’s history and physical (HP) file. However, he also acknowledged that these data is not easy to find in the file because there is no standardized protocol or form for collecting this kind of data.

The History and Physical file is where you might find it (police involved shooting information). The history file states whether the patient is a victim of a gunshot wound (mechanism of injury), the source of the gunshot which will indicate whether it was a domestic dispute or the police. However, most HPs (History and Physical file) only have the mechanism of injury but that does not give any information on context. But the real question is, is this data easy to get in the file? The answer to that is no! But it is very easy to create a database to collect this kind of information to answer this question.

Dr. Smith explained in detail some of the ED’s logistical issues with collecting police shooting data:

We don’t get an accurate history because we do not get that information right away when the person is brought into trauma. Some people (staff) may write in the HP that the patient was shot by the police while others may not. The circumstances of the injury are not consistently logged in the HP. We do not systematically collect this data because we are trained to take care of patients not to collect that kind of data. But if a patient survives he or she should be interviewed by the VIP (Violence Intervention Program) on who shot you. This is why we need a VIP (Violence Intervention Program) social worker to ask those questions.

Kim, the violence prevention and trauma outreach nurse, described the lack of standardization in the collection of violent injury data by the trauma staff.

The information is not standardized. Trauma and EMS are only concerned about the mechanism of injury so it’s difficult to get a reliable account of what happened. Even if we pulled patients that were gunshot victims and go into their chart and read the notes, what data are we going to extract to get the context of injury? One complaint I could see is that ER staff may differ on the way they describe the injury. One person might say ‘arm injury’ but it could be a gunshot wound.

Despite some skepticism on the logistics challenges in the data collection, Kim was optimistic that the data could be collected in a more standardized format:

The nurses could survey each patient with a GSW (gunshot wound) to determine if it was a police involved shooting. The ER and trauma centers provide the best opportunity to get the context of gunshot victims.

Daphne (trauma nurse) also supported a more standardized approach for data collection:

There should be some standardized way to collect that information using a check box mechanism.

Carlton discusses solutions for collecting police-involved shooting data. He proposed one novel approach to the data collection suggesting that researchers should be situated in the ED 24 hours a day to collect this information. While this approach may be costly, it could provide an innovative and more accurate approach to collect data.

Right now there is no standardized procedure for collecting police involved shooting data. But the hospital can ask if the shooting was by a stranger or the police. If a person gets brought in by the police at that point medical staff can ask a simple question: Was it community or individual-level interpersonal violence or was it by the police? It is a simple dichotomous yes or no answer. What was the context of your injury and leave it at that. Trauma staff has the capacity to ask routine questions about who shot you. If the patient is able to speak hospital staff should ask, who shot you? I mean we ask questions on substance use when we use the SBIRT (Screening Brief Intervention Referral for Treatment). How much energy does it take to add another question to that form? Was your injury the result of someone in the community or law enforcement? Busy trauma units like here where we see a lot of gunshot wound victims there should be a 24 hour research staff in the trauma unit that screens for patients with gunshot wounds.

Dr. Smith discussed other innovative ways that the data could be collected through the use of social workers affiliated with the HVIP:

We should have screening questions for who was the perpetrator. The way I would do it, I would want to know, did you know the perpetrator? Was it a person in your family, someone in your neighborhood or law enforcement? All of the questions would have to be under the guise of being health-related, that we are trying to stop violence and recidivism. We could then ask questions and create a database that collects data on the circumstances surrounding the injury. Most importantly this has to be done by a VIP social worker that has gained the patient’s trust because patients will be not as forthcoming with doctors. I don’t ask them who shot you because the majority of the information they give me is not forthcoming.

Here, Daphne also describes some of the challenges/barriers trauma nurses may face collecting data. Daphne suggests that trauma nurses affiliated with the HVIP could be a more effective resource for collecting police-involved shooting data:

There are some nurses that are case managers who act in the capacity of a social worker, who could obtain data if the patient is admitted to the trauma floor. When I was a trauma outreach nurse part of my job was to identify the circumstances of the injury. But as part of that process, the nurse has to make sure they are not getting specific details about what happened. It can be just a yes or no answer. As a violence prevention and trauma outreach nurse I was allowed to probe into the circumstances of the injury, but a regular nurse may not be allowed to do it or she may be hesitant to do it. Nurses not involved in a HVIP may not ask those questions. Also an outreach nurse is also not at the hospital 24 hours a day and not on weekends. So someone has to be there to collect the data at all times. But a hospital could create a HVIP position where someone could ask patients those questions.

Once the data is collected, the looming question remained: How will the HVIP determine the accuracy of the data collected? Carlton (HVIP co-director) suggested that one approach could compare the data collected by the hospital to data collected by the police. This strategy parallels the procedures implemented in the UK 13:

This way the hospital could have a separate database that the police are not privy too and that is not controlled by law enforcement. The hospital data could be compared to the police data.

One emerging theme consistently mentioned by staff during the interviews was the potential to collect community violence and police-involved shooting data from EMS workers.27 Here, Daphne discusses how this data could be collected:

Trauma could gather additional information on the context of injury from the EMS reports.

Other trauma staff also emphasized the need to use multiple sources for data collection:

We need three sources of data that do not include the police. We need patient, medical staff and the EMS reports. When the EMS comes into the hospital we have the opportunity to potentially interview them to determine whether it was a police involved shooting.-Carlton

Dr. Smith emphasized the importance of utilizing EMS data to capture the context of the injury:

The earliest documentation of what happened is really done by the EMS. The only way you can get a more accurate account of what happened is from the EMS runsheet because they (EMS) type that kind of information in such as where the victim was found, mechanism of injury, perpetrator, etc. We can pull the runsheets to see if they were police-involved. However, there is no field for police involved shootings on the runsheets, so we have to look physically at the runsheets and read through them because there is no method to enter that data into trauma outcomes. There is no field for who was the perpetrator.

One logistical theme discussed by Daphne which is a significant barrier to the data collection was the presence of the police at the hospital. Unless patients were shot accidentally, patients injured via police-involved shootings were placed under arrest. The patient’s legal status makes it difficult for researchers to access this population. However, Daphne noted that trauma nurses and doctors have unfettered access to patients. Despite the police being present in the room of the patient, Daphne was asked whether it was possible for trauma staff to ask questions surrounding the shooting and if patients would be forthright or withhold information, fearing reprisal from the police:

In my last job as a violence prevention and trauma outreach nurse I could ask questions about how the patient was shot right in front of the police. Some were reluctant to talk if they were doing something they should not have been doing. But that is something I need to look into regarding whether you can do that here at this hospital because each hospital is different. As a trauma nurse, I am not sure whether the police can prevent a patient from being asked those questions if the patient is under arrest.

Here, Dr. Smith describes how some patients are willing to disclose this information with the police present in the room:

Guys don’t want to give up the information if someone else shot them, but if the police did it they will give it up. I never had a patient shot by the police not tell me they were shot by the police.


Trauma staff overwhelmingly agreed that the hospital has an ethical responsibility to record and report police-involved shootings. While some disagreed on the attitudinal, logistical, and the legal ramifications associated with collecting this kind of data, most reported that the hospital had the ethical responsibility to be responsive to the community to record and report police-involved shootings. Here, Daphne explains why:

Collecting this data is important for the sole purpose of violence prevention regardless of who the perpetrator is. For the community this data is needed. We have to look outside the walls of the hospital. We have to serve the entire community.

Carlton also described the ethical responsibility of the hospital to record and report police shootings:

If we are really concerned about injuries caused by law enforcement we would allow medical staff to ask questions. If we find out how many injuries are the result of legal intervention it gives us an impetus for how we can control this trend. But if we rely solely on police data we are never going to address it. If we conceptualize violence/trauma as a disease it should be treated as a disease and we should know the epidemiology of it regardless of the source even if it is the police.

Dr. Smith expressed mixed feelings regarding the hospital’s ethical responsibility to collect and report this data:

I think we (hospital) should focus on the numbers (police shootings) but we should know all the numbers. I think it’s fair to know all the numbers including the numbers of Black men shot by each other. I mean it’s great to protest, do marches, Black Lives Matter, and Hands Up demonstrations in our communities but it has to be equal. What about doing that for the 108 homicides in Baltimore thus far this year? If we are going to intervene let’s intervene on that.

While the majority of trauma staff agreed that the hospital has an ethical responsibility to record and report police-involved shootings, some staff advocated that hospitals are responsible for reporting all persons involved in shootings regardless of whether they were law enforcement-related or interpersonal. HVIPs have the potential to do both.


The legal ramifications associated with collecting police shooting data were the most consistently discussed issues among staff. Trauma staff nurses expressed reluctance in delving too far into the circumstances of the shooting, fearing that they may be in violation of the law. Here, Daphne discusses her trepidation:

I don’t know if nurses will ask questions regarding who shot you because once you start asking questions the police can call you (nurse) into court. Especially if the information nurses are collecting in their notes may get them subpoenaed. Some nurses may be concerned about being a witness in court so they may feel hesitant to document or identify police involved shootings because of the legal ramifications that might occur. I don’t really know what the hospital is legally able to do in such cases.

However, Daphne also provided a counter-narrative that a trauma nurse assigned to a HVIP and a regular nurse have different responsibilities in regard to collecting police-involved shooting data from patients:

Trauma nurses are told not to probe patients and not to ask specific questions unless they are related to a domestic (dispute). As a violence prevention and outreach nurse you can probe but a regular nurse cannot. Nurses who are not affiliated with an HVIP cannot ask those kinds of questions.

Dr. Smith expressed apprehension that the data collected by trauma staff would not be permissible in court. According to Dr. Smith, because information collected by trauma staff may not be accurate, legally it may be treated as unreliable data in court:

If hospitals get this information (police involved shooting data) how do we know it’s true? This is politically charged because you can’t just say this happened in court, you have to prove it happened. When I go to court I have to depend on the EMS report that’s all I can go on about what happened. If a patient survives you can interview them about what happened, but the only thing the hospital can say is this is the information they gathered for the report, they do not know if it is factual. So we cannot ask legal questions that would be used for court because you wouldn’t want that subpoenaed. We also cannot ask whether they were doing anything illegal at the time they were shot, but we do need to get at what they were up to in their neighborhood when they were shot.

Because some patients may be reluctant to divulge police-involved shooting data in fear of police reprisals, Dr. Smith proposed novel ways to circumvent the legal obstacles of collecting police-involved shooting data from patients by framing questions as health-related:

We have to tell to the patient this is not a police report. This is a health report on the circumstances surrounding your shooting no information will be given to the police.

While the sharing of data between the police and the hospitals has worked well in the UK, some trauma staff raised concerns on whether law enforcement in USA would be willing to engage in a data sharing relationship with the hospital to record community violence.

The data would serve as a foundation if people see value in it. But law enforcement will see no value in it because it will make them look bad. But they should want to know so they can figure out alternative ways in dealing with how police are trained to negotiate violence. This information will help law enforcement to better train officers for rules of engagement with violence and how to better deal with it. But they (law enforcement) don’t want it.-Carlton (Co-Director HVIP)

Dr. Smith summed up his feelings about the tenuous relationship between the hospital and law enforcement:

It is a complex issue with many moving parts. I mean hell, how do you get at all these sociological issues when the only story that you can really rely at the moment is the police?

However, emergency physicians who treat victims of violent assaults should not assume that only the police know about all or most violence.12


To our knowledge, there are no studies on how hospital violence intervention programs in USA can record and report police-involved shootings. The qualitative data presented in this article suggest that HVIPs have the potential to collect reliable police-involved shooting data. Education and training HVIP staff is needed to provide innovative approaches for the collection of these data. Many HVIPs collect data on the circumstances of injury; these data are of particular importance in identifying the risk factors for repeat violent injury because the data is used to develop effective violence intervention strategies.2832 Yet, to our knowledge, there are no studies on how HVIPs collect police-involved shooting data. Furthermore, the vast majority of studies on recording community violence by medical and police services have been conducted in the UK. Even those studies do not acknowledge police brutality as a factor in violent assaults treated by the ED. An exhaustive review of the literature also found few studies on police abuse as a public health concern.3,33 From the perspective of ED physicians, the data are scant on how emergency physicians perceive and report excessive use of force by the police.1

Our findings suggest that there are several strategies HVIPs can use to collect reliable police-involved shooting data. One important resource are the first responders and the EMS reports which provide the location of where the victim was found, mechanism of injury, and circumstances surrounding the injury. One ED physician in our study noted that this information can be extracted from the runsheets generated by the EMS. However, one limitation of the runsheets is that they do not include a field for police-involved shootings. Consequently, ED staff must physically read through each runsheet to determine the circumstances of the injury. The omission of a field for police-involved shootings may suggest that EMS and the health sector, at least as a matter of protocol, do not acknowledge police abuse as a potential factor in violent victimization and injury. Therefore, in order to ascertain data on the context of police-involved shootings, future studies should assign a researcher(s) to the trauma unit to screen all EMS workers for firearm-related shootings. An additional procedural and format change to the runsheets should include a field that captures officer-involved shootings. This field would be standardized on all EMS runsheets. By standardizing the collection of this data, EMS staff would be required to record and report police-involved shootings, eliminating the potential for these incidents to be underreported or not accurately reported in their written notes on the runsheet. This format would lessen the threat of reporting errors.

In addition, EMS reports and HP files should include a field for documenting police-involved shootings. Each report and file should have a field that captures the circumstances of the shooting (e.g., relationship to perpetrator, stranger, someone in community, family member, law enforcement). Recording and reporting this should not be left to the discretion of the staff member to document this information. EMS and trauma staff should be educated and trained to collect this data. HVIPs must also develop a software which will allow them to share and exchange data. Harmonization of police and hospital injury datasets is also needed to measure police-involved shootings.

HVIPs can begin collecting this data using a standardized and systematic approach. One novel approach is the inclusion of questions on screening tools completed by victims of firearm-related shootings that directly ask whether the shooting was police-involved. The screening, brief intervention, and referral to treatment (SBIRT) is used by primary care centers, hospital emergency rooms, and trauma centers to quickly screen and assess the severity of substance use and identifies the appropriate level of treatment. SBIRT consists of three components: (1) structured assessment; (2) brief intervention; and (3) referral to services. Referral to treatment provides those identified as needing more extensive treatment with access to specialty care. SBIRT offers an opportunity to identify problem drinking and substance abuse and trigger intervention. Because drug and alcohol intoxication are significantly correlated to violent injury, the SBIRT should be utilized as a screening tool for all victims of violent injury. Included in the SBIRT should be questions regarding the contextual factors surrounding the injury. One question should investigate the relationship between the perpetrator and victim of the injury (e.g., stranger, acquaintance, law enforcement). Contextual data on the factors associated with the injury should also be used to determine whether a victim of violent injury should be referred to an HVIP for services.

The lack of a standardized and systematic method for collecting police shooting data remains problematic. Based on the data we collected from trauma staff at PGHTC and a review of protocols used by HVIPs in two other cities which assess risk factors for violent injury, we found no questions on their surveys which directly asked the respondents whether they were victims of excessive force used by the police. It is worth noting that in our longitudinal ethnographic pilot study at PGHTC on risk factors for recurrent violence, we did not ask participants questions regarding police brutality in our first two waves of interviews. Only in the final wave 3 interviews did we include in our instrument questions that probed for incidents of police brutality. Our research team neglected to acknowledge police brutality as a public health concern among Black male victims of violent injury. In our interviews, we asked in-depth detailed questions on the circumstances and context of each injury but we failed to ask whether any injuries were related to police violence. The recent high-profile police-involved shootings among Black men forced our research team to confront this issue in our data collection.

Another important approach for collecting data is the social workers and trauma nurses assigned to HVIPs. Like EMS workers, social workers and violence prevention and outreach trauma nurses assigned to HVIPs play a critical and valuable role in the collection of data. These members of HVIPs have access to patients involved in police shootings who may be under arrest while they are being treated for their injury. Because populations that have experienced police-perpetrated abuse may fear that police might exacerbate violence or further traumatize victims, HVIP social workers and nurses can reduce fear, ensure safety and confidentiality, and develop trust and rapport with patients.3 As one trauma nurse mentioned, HVIP nurses have the permission to probe for circumstances surrounding the injury where other trauma staff cannot.

Unlike trauma surgeons, HVIP social workers and nurses also have constant and continuous interaction with victims of violent injury while they are being treated in the hospital setting. These supportive relationships extend beyond the hospital and continue after the patient is discharged. HVIP social workers and outreach nurses develop and facilitate violence intervention treatment plans that include wraparound psychosocial services, employment/educational training referrals, and collaboration with departments of probation and parole if the patient is under criminal justice supervision.

In order to successfully collect this data, hospital administrations must also educate and train HVIP staff on the legal implications related to gathering police-involved shooting data. Trauma surgeons and nurses expressed a reluctance to collect data because they feared being subpoenaed. One nurse reported that she was unaware of the hospital’s policies on the collection of police-involved shooting data. If more clarity was given to trauma staff regarding how sensitive information would be handled by the police and courts, it may reduce the fear that many trauma staff encounter when deciding whether to probe on the circumstances surrounding the injury. This also requires a change in the traditional hospital culture to shifting their mission as simply care providers to an ethical commitment to reduce community violence even when the source of violence is law enforcement. While there may be a significant pushback from hospitals, HVIPs are in a unique position to address this politically charged issue. As one trauma physician noted, HVIPs can position their concern as a health-related issue and not a police issue.

Although our study sample was small, our findings support previous research findings on emergency physicians’ perceptions of excessive force that these events should be reportable. More importantly, EPs should advocate for the patient when suspected excessive force is an issue without jeopardizing their working relationship with local, state, and federal law enforcement agencies.1,12,19

We recommend that public health policies and research address the prevalence, nature, and public health implications of police violence.3 There has been little or no requirement for ED staff to collect information on police-involved shootings. Although legislative steps have been taken to protect society from communicable diseases, public health policies have rarely been developed and enforced to protect society from police violence. There is a lack of a legal framework to ensure that ED physicians and HVIPs can report excessive police force. This may be an obstacle to protecting patients and communities. Legislation should protect ED physicians and HVIPs in reporting police violence and abuse. Legislation is also needed to protect citizens and communities particularly low-income communities of color which are disproportionately impacted by excessive police force, police abuse, and police-involved shootings.

In the UK, the Crime and Disorder Act of 1998 identified health authorities as partners with whom local government and the police must work collectively to address violence and crime. The Cardiff Violence Prevention Programme model in the UK is a novel approach to using relevant available data and expertise from health and law enforcement officials to reduce violence and to develop violence prevention strategies.34 Because USA has far more police-involved shootings than any other industrialized country in the world, there must be a federal legislation which requires hospitals to record and report all police-involved shootings to their state’s health department. These data could be compared to police data submitted to the state to measure the accuracy of reporting. Hospital data would also be submitted to the CDC or a national health statistics database. Studies have shown that EDs report 25–50 % higher on the number of violent assaults recorded by the police and that only 4 % of law enforcement agencies provide police shooting data to the FBI.

Federal legislation requiring that hospitals record and report police-involved shooting data could address the nature and prevalence of police-involved shootings. Partnerships between hospitals and law enforcement could then identify strategies that reduce police violence.3 EDs particularly HVIPs situated in Level I and Level II trauma centers are in a unique position to partner with law enforcement on reducing police violence particularly because EDs and HVIPs evaluate more patients with police-citizen encounters than any other medical specialty.1

Do EDs have an ethical obligation to report excessive use of police force and police violence? Overwhelmingly, the data collected from ED physicians on their perceptions of police violence suggest that hospitals and physicians have a responsibility to report events. HVIPs have an ethical commitment to reduce violence and repeat violent injury among the patients they serve. This commitment also includes reducing police violence. The consensus among participants in our small sample suggests that HVIPs should collect this data because hospitals must be responsive to the needs of the patient, the community, and the public services that serve them. While we agree that domestic violence should be mandatorily recorded and reported, EDs and HVIPs have an obligation to report all forms of violence. One ethical concern that has been raised is the issue of confidentiality. We argue that the question of confidentiality could be minimized as all data could be anonymized.13,15


This research involved interviews and informal conversations with trauma staff as part of a larger study on violence and trauma among young Black men; therefore, we did not specifically focus on the role and function of HVIPs in our data collection. As a result, we collected data from one emerging HVIP and five trauma-affiliated staff. This limits the generalization of the data. There are 36 HVIP programs in the National Network of Hospital-Based Violence Intervention Programs. For the data collection, we utilized a convenience sampling strategy; we chose to interview staff from an emerging HVIP situated in the same hospital where the larger research study was conducted.

Emerging HVIPs are defined as programs still in development. As an emerging HVIP, this program did not provide patients with services or referrals. The HVIP was still in the process of developing its mission, staff, and long-term funding for sustainability. In the future, we recommend the inclusion of a larger sample of HVIPs on their perceptions and challenges regarding the collection of police-involved shooting data. We also did not include hospital administrators or EMS workers in our sample. Both could have provided richer data for our analysis and discussion. It is also possible that the questions we asked did not effectively assess ED and HVIP staff perceptions. We asked respondents to make subjective judgments without objective evidence. For example, trauma surgeons and nurses in this study were unaware that law enforcement was not mandated to report police shooting data to the Justice Department. Knowledge of this data may have changed their perceptions and responses.

Future Research

More evaluations are needed that compare community violence recorded and reported by hospital EDs to police data. Future studies should investigate the ways hospitals and police record and report community violence. Presently, there are no qualitative studies on the perceptions and challenges of HVIPs regarding recording and reporting police violence and excessive use of force. The majority of studies we found that compare hospital violent assault data to police data have been conducted in the UK. USA lags far behind on this issue. In the USA, the comparison and evaluation of hospital to police data is scant. Studies on police-involved shootings should also examine EMS data. As first responders, EMS workers are the first to arrive on the scene and are in a unique position to collect data at the earliest point of the event. From a health data collection approach, they are furthest upstream. Finally, HVIPs are uniquely positioned to conduct studies on police violence. HVIPs have access to trauma and EMS staff, police officers, and patients to collect data. These are four valuable sources of data. HVIPs are also in strategic position to screen firearm-related injuries 24 hours a day.


Only 4 % of law enforcement agencies report police-involved shootings to the Justice Department. At the time this article was written in July 2015, an independent Washington Post analysis of police-involved killings found that 500 people were killed by the police since the beginning of the year.35 This number exceeds the average of 400 “justifiable homicides” reported by law enforcement to the FBI every year. EDs reporting of violent assaults have found wide disparities in their recorded data and the data recorded by the police. Despite the World Health Organization’s (WHO) classification of police officers’ excessive use of force as a form of violence, public health researchers have produced scant research characterizing police-perpetrated abuse and its significance for public health.3

Evidence regarding police shooting data suggests that it is grossly underreported by the police, the Justice Department, and the CDC. Recently, legislation (Safe Justice Act of 2015; Police Reporting Information Data and Evidence PRIDE Act of 2015) has been proposed by members of Congress to create a national database for police-involved shootings. Both bills would require mandatory reporting of police-involved shootings to state law enforcement and the mandatory reporting of state level data to the federal government. What is problematic about the proposed legislative bills and previous legislation such as the DCRA is that the bills do not include funding for the design and collection of data. For example, reports produced by Florida state law enforcement do not conform to FBI requirements. Funding for the standardization of the design and collection of data would need to be provided to 17,000 law enforcement agencies. Furthermore, approximately 80 % of all police departments have less than 10 officers. Most departments do not have staff specifically assigned to collect statistical data. The funding for the design and collection of data being absent, these bills are essentially “unfunded mandates.” As a result, these bills are unlikely to solve the problem regarding accurate data collection of police-involved shootings.10

Thus, novel innovative approaches are needed to capture reliable and accurate police-involved shooting data. HVIPs provide a unique opportunity to document the prevalence of police violence and abuse. As researchers who study violence and trauma among young Black men using the ED as a clinical setting, we strongly advocate that EDs and HVIPs systematically identify, monitor, and surveil all firearm-related injuries. We believe this is a rational step to improve recording, reporting, and preventing police violence. Answering the question, “Who shot you?” is the first step toward addressing how the health sector and law enforcement can work together to improve the health of the individuals and communities they have taken an oath to “protect and serve.”

Contributor Information

Joseph B. Richardson, Jr, ude.dmu@5rahcirj.

Christopher St. Vil, ude.olaffub@livtsc.

Carnell Cooper, ude.mmu@repoocc, gro.htlaehsnoisnemid@repooc.llenrac.


1. Hutson H, Anglin D, Rice P, Kyriacou D, Guirguis M, Strote J. Excessive use of force by police: a survey of academic emergency physicians. Emerg Med J. 2009; 26(1): 20–2. [PubMed]
2. Krug E, Dahlberg L, Mercy J, Zwi A, Lozano R. World report on violence and health. Geneva: World Health Organization; 2002.
3. Cooper H, Moore L, Gruskin S, Krieger N. Characterizing perceived police violence: implications for public health. Am J Public Health. 2004;94(7):1109–1118. doi: 10.2105/AJPH.94.7.1109. [PubMed] [Cross Ref]
4. Lowery W. How many police shootings a year? No one knows. Washington Post. September 8, 2014. Retrieved from
5. Schmidt M. Scant data frustrates efforts to assess number of shootings by police. NY Times. April 8, 2015. Retrieved from
6. Federal Bureau of Investigation. Uniform crime report: Crime in the United States 2013. U.S. Department of Justice. 2014. Retrieved 1 June 2015, from
7. Barry R, Jones C. Hundreds of police killings are uncounted in federal stats. Wall Street J. 2014. Retrieved 1 June 2015, from
8. Gabrielson R, Grochowski Jones R, Sagara E. Deadly force, in black and white. ProPublica. 2014. Retrieved 1 June 2015, from
9. Bratton W. New York City Police Department annual firearms discharge report 2013. New York: New York City Police Department; 2014.
10. Graham DA. Why is it so hard to track police killings? The Atlantic. June 3, 2015. Retrieved from
11. Hargarten S, Waeckerle J. Docs and cops: a collaborating or colliding partnership? Ann Emerg Med. 2001;38(4):438–440. doi: 10.1067/mem.2001.118615. [PubMed] [Cross Ref]
12. Shepherd J. Emergency medicine and police collaboration to prevent community violence. Ann Emerg Med. 2001;38(4):430–437. doi: 10.1067/mem.2001.114317. [PubMed] [Cross Ref]
13. Sutherland I, Sivarajasingam V, Shepherd J. Recording of community violence by medical and police services. Inj Prev. 2002;8(3):246–247. doi: 10.1136/ip.8.3.246. [PMC free article] [PubMed] [Cross Ref]
14. Howe A, Crilly M. Identification and characteristics of victims of violence identified by emergency physicians, triage nurses, and the police. Inj Prev. 2002;8(4):321–323. doi: 10.1136/ip.8.4.321. [PMC free article] [PubMed] [Cross Ref]
15. Florence C, Shepherd J, Brennan I, Simon T. Effectiveness of anonymised information sharing and use in health service, police, and local government partnership for preventing violence related injury: experimental study and time series analysis. BMJ. 2011;342(16 2):d3313–d3313. doi: 10.1136/bmj.d3313. [PubMed] [Cross Ref]
16. Kellermann A, Bartolomeos K, Fuqua-Whitley D, Sampson T, Parramore C. Community-level firearm injury surveillance: local data for local action. Ann Emerg Med. 2001;38(4):423–429. doi: 10.1067/mem.2001.117273. [PubMed] [Cross Ref]
17. Kellermann A, Rivara F, Lee R, Banton J, Cummings P, Hackman B, Somes G. Injuries due to firearms in three cities. N Engl J Med. 1996;335(19):1438–1444. doi: 10.1056/NEJM199611073351906. [PubMed] [Cross Ref]
18. Cole T, Flanigan A. What can we do about violence? JAMA. 1999;282(5):481. doi: 10.1001/jama.282.5.481. [PubMed] [Cross Ref]
19. Shepherd J, Lisles C. Towards multi-agency violence prevention and victim support: an investigation of police-accident and emergency service liaison. Br J Criminol. 1998;38(3):351–370. doi: 10.1093/oxfordjournals.bjc.a014252. [Cross Ref]
20. Rich J, Grey C. Pathways to recurrent trauma among young black men: traumatic stress, substance use, and the “Code of the street” Am J Public Health. 2005;95(5):816–824. doi: 10.2105/AJPH.2004.044560. [PubMed] [Cross Ref]
21. Purtle J, Rich L, Bloom S, Rich J, Corbin T. Cost−benefit analysis simulation of a hospital-based violence intervention program. Am J Prev Med. 2015;48(2):162–169. doi: 10.1016/j.amepre.2014.08.030. [PubMed] [Cross Ref]
22. Chong V, Smith R, Garcia A, Lee W, Ashley L, Marks A, et al. Hospital-centered violence intervention programs: a cost-effectiveness analysis. Am J Surg. 2015;209(4):597–603. doi: 10.1016/j.amjsurg.2014.11.003. [PubMed] [Cross Ref]
23. Cooper C, Eslinger D, Stolley P. Hospital-based violence intervention programs work. J Trauma Inj Infect Crit Care. 2006;61(3):534–540. doi: 10.1097/01.ta.0000236576.81860.8c. [PubMed] [Cross Ref]
24. Rich J. Wrong place, wrong time : trauma and violence in the lives of young black men. Baltimore: Johns Hopkins University Press; 2009.
25. Lim H, McCart M, Davies W, Calhoun A, Melzer-Lange M. Risk for repeat emergency department visits for violent injuries in youth firearm victims. Clin Med Trauma Intensive Med. 2009;2:1–7.
26. Glaser B, Strauss A. The discovery of grounded theory. Chicago: Aldine Pub. Co.; 1967.
27. Ariel B, Weinborn C, Boyle A. Can routinely collected ambulance data about assaults contribute to reduction in community violence? Emerg Med J. 2013;32(4):308–313. doi: 10.1136/emermed-2013-203133. [PubMed] [Cross Ref]
28. Cooper C, Eslinger D, Nash D, Al Zawahri J, Stolley P. Repeat victims of violence. Arch Surg. 2000;135(7):837. doi: 10.1001/archsurg.135.7.837. [PubMed] [Cross Ref]
29. Claassen C, Larkin G, Hodges G, Field C. Criminal correlates of injury-related emergency department recidivism. J Emerg Med. 2007;32(2):141–147. doi: 10.1016/j.jemermed.2006.05.041. [PubMed] [Cross Ref]
30. Davis J, Castilla D, Schulman C, Perez E, Neville H, Sola J. Twenty years of pediatric gunshot wounds: an urban trauma center’s experience. J Surg Res. 2013;184(1):556–560. doi: 10.1016/j.jss.2012.12.047. [PubMed] [Cross Ref]
31. Liebschutz J, Schwartz S, Hoyte J, Conoscenti L, Christian A, Muhammad L, et al. A chasm between injury and care: experiences of black male victims of violence. J Trauma Inj Infect Crit Care. 2010;69(6):1372–1378. doi: 10.1097/TA.0b013e3181e74fcf. [PMC free article] [PubMed] [Cross Ref]
32. Keough V, Lanuza D, Jennrich J, Gulanick M, Holm K. Characteristics of the trauma recidivist: an exploratory descriptive study. J Emerg Nurs. 2001;27(4):340–346. doi: 10.1067/men.2001.116214. [PubMed] [Cross Ref]
33. Cooper H. War on drugs policing and police brutality. Subst Use Misuse. 2015; 50(8–9): 1188–1194. [PMC free article] [PubMed]
34. Boyle A, Snelling K, White L, Ariel B, Ashelford L. External validation of the Cardiff model of information sharing to reduce community violence: natural experiment. Emerg Med J. 2012;30(12):1020–1023. doi: 10.1136/emermed-2012-201898. [PubMed] [Cross Ref]
35. Kindy K. Fatal police shootings in 2015 approaching 400 nationwide. Washington Post. May 30, 2015. Retrieved from

Articles from Journal of Urban Health : Bulletin of the New York Academy of Medicine are provided here courtesy of New York Academy of Medicine