How We Reported on Gunshot Victims’ Access to Trauma Care

We mapped 12,000 gunshots to reveal a hole in New York City’s trauma system, and other newsrooms can do the same in their areas.

(The Trace)

In October we published a story showing that gunshot victims in New York City are more likely to die the further they are from a trauma center. An area of southern Queens, where fatality rates are particularly high, seems to be the most affected by this phenomena. For all of the residents in this region, there is only one trauma center to treat the most severe injuries, and it is financially struggling to keep its doors open.

We mapped more than 12,000 shootings recorded by the NYPD over a 9-year-period to draw our conclusions, and examined how the number of intensive care beds in nearby trauma centers might influence survival rates. You can read our detailed analysis here.

Within a week of its publication, the story had caught the attention of local politicians, and they pledged to address the lack of trauma care in southern Queens. Trauma care and gun violence researchers have praised the quality of the data analysis, and say the story shines a light on an important health disparity.

When the causes of harm contain multiple factors, it often takes a complex approach to data analysis to draw a direct connection for readers, researchers, and policy makers. And our story is a prime example of how this can be done.

The project started with an innovative collaboration between The Trace, a nonprofit newsroom covering gun violence, and Measure of America, a data and research team at the nonprofit Social Science Research Council that had recently hired a journalist courtesy of a grant from the Helmsley Charitable Trust. We were brought together because of our beliefs in reporting centered on data and evidence-based analysis. Measure of America was also starting to focus on supplying newsrooms in New York City with tools to enhance their beat reporting and was looking to focus more on gun violence. While The Trace uses data to support its reporting, Measure of America uses its social science expertise to reveal new information found in datasets.

Our two groups recognized how we could build on our specialties to produce a project with greater impact than either of our groups working separately. Once The Trace received some long-awaited data from the NYPD, a project centered in its analysis seemed like the perfect fit for our skills and interests.

The three of us dove into the data and reporting, with assistance from many others on both teams. Sean led the project, Maya had her feet in the data and reporting, and Laura headed up the statistical analysis. We wrote the analysis whitepaper and story together, and the piece was edited and co-published by THE CITY, a local newsroom.

The story added to findings by researchers across the country that gunshot fatality is closely tied to trauma center access. Want to check out whether this is happening in your city? Read on for an outline of what we did and find out how you can do the same.

Story Overview

After nearly two decades of declining firearm homicide rates in the United States, the most recent data available from the CDC suggests that gun homicide rates are increasing to levels not seen since 1998. Most of this violence is concentrated in urban centers and large cities. Increasingly, experts are considering how access to health care and gun violence overlap.

Previous research in Detroit, Chicago, and LA has shown that the further gunshot victims are from trauma centers, the more likely they are to die. Our reporting and analysis based in New York City builds on this research, finding similar results.

To get started, your newsroom can check the location of trauma centers in your city that have been verified by the American College of Surgeons (ACS). Then, you’ll want to start homing in on where these trauma centers are located in relation to areas of concentrated gun violence. As a rough approximation of where shootings are happening in any given location, The Trace has mapped every shooting incident in the Gun Violence Archive from 2014 to 2018. You’ll be able to get official data through the local police department, or even the city’s open data portal (more on that, below). And The Trace has raw shooting data from 56 police and sheriff’s departments that was obtained directly from police departments available for download.

With year-over-year rises in the firearm death rate in the U.S., it is becoming increasingly important to understand what factors might be contributing to people losing their lives from gunshots. This city-level reporting plan can help local newsrooms tease apart one of these factors, and has the potential to draw attention to areas that have been affected by a lack of access to trauma care and/or trauma center closings.

Data You’ll Need

  • Data from the local Police Department that includes all shootings, both nonfatal and fatal. Ensure that there is a large enough sample size to be able to draw significant conclusions. (The Trace has this data through 2017 for nearly two dozen major U.S. cities, and some of the necessary data for others available free of charge.) The data fields needed are:
    • Whether the shooting incident was fatal or nonfatal
    • Date of the incident
    • Physical location of the incident in either block-level address or latitude and longitude
    • Crime-type of the incident (e.g. homicide, suicide, justified shooting, assault, etc.)
    • Other useful fields:
      • Circumstances (e.g. domestic violence, etc.)
      • Victim demographics (i.e. race/ethnicity, age, sex)
      • Bullet caliber
      • Injury location (i.e. where the victim was shot)
      • Whether the shooting happened indoors or outside
  • List of trauma centers in the area
    • American College of Surgeons: the ACS allows users to search for “verified” trauma centers by city and zip code. An important element to note is that ACS verification is voluntary. While it’s considered by many to be the gold-standard for trauma center evaluation, state designations may differ.
    • State-level trauma designations: State departments of health will often have this information on-hand, and in many cases the hospitals can be found online. Ideally, the list you obtain will indicate whether the facility is a level I trauma center, or whether it offers any other specialty services. As an example, New York’s repository is available here.
  • List of trauma center closures over the timeframe for your data.
    • You can request this (informally) from your state department of health.
    • You can also read through previous media stories and state health reports to identify when hospitals opened and closed.
    • Using the Internet Archive’s Wayback Machine to dig through time on your local health department’s website can be very useful in identifying when trauma centers opened, closed, or changed designations.
  • Street-grid mapping data of the city
    • The U.S. Census has files for geographic features, including road networks (link1, link2)
    • Many cities with online data portals host downloadable files for their street networks. For instance, here is the street grid for New York City.
  • Analysis Tools

    • QGIS: A free mapping software.

    • Google Sheets

    • A reliable coding language such as R or Python

    Analysis Methods

    In QGIS, use the Service Area From Layer tool in the Network Analysis toolbox to generate polygons showing the area within a given distance along the street grid from each trauma center. For example, generate 1-, 2-, and 3-mile regions around each trauma center. Adjust the specific distances as appropriate for the scale of your city and the number of trauma centers.

    Then, use a spatial join or the Select by Location tool to categorize each shooting based on which distance band it falls into (i.e. less than one mile, one to two miles, two to three miles, three or more miles). Calculate the fatality rate of the shootings that occurred in each of these bands by dividing the number of fatal shootings by the total number of shootings in the zone.

    Researchers Giovanni Circo and Andrew Wheeler have a package of coding files to recreate the analysis that they performed in Philadelphia. The methods they used are more advanced, but are easily adaptable for your city with some coding expertise. You can download the files here.

    Reporting Plan

    • After you’ve run through the steps in our analysis, look for local reports on the trauma system in your city. Your state and city departments of health are a good place to start. As an example, in New York, the State Department of Health prepares bi-annual reports on the trauma system, and the NYC Department of Health produces community health reports.

    • Check for recent closures, and any closure plans the hospital produced. The methods used in our analysis can potentially point to areas that have seen a disproportionate rise in gunshot wound deaths due to trauma center closures. If trauma center closings have been a problem in your area, there is a chance that community groups have been trying to get attention for the issue, or have maybe testified before state lawmakers.

    • Locate survivors in the areas most heavily affected; media reports can help with this. In our experience, survivors of gun violence often feel the effects of trauma, even if they were not directly involved in the incident. Aside from the loss of life, gun homicides can leave lingering pain in the community.

    • Speak with state and city trauma officials. These insiders may have access to industry reports that haven’t been publicized, and that explain why there is a lack of trauma coverage in the areas today. They might also point you in the direction of trauma holes to look into.

    • If you find an area that appears to suffer from a lack of access to trauma care, start digging into the hospitals that are providing coverage for the area. There may be lawsuits for inadequate care or suits that have been brought forward by whistleblowers. If a trauma center is overburdened, it’s likely going to show in areas aside from patient mortality. ProPublica has an ER Wait Watcher that contains data on how long it takes patients to be seen by a doctor. Medicare.gov rates thousands of hospitals based on their quality of care and the Association of Health Care Journalists (AHCJ) allows users to search through citation reports from federal inspections. The non-profit Investigative Reporters and Editors (IRE) also has a number of tipsheets available for how to look into your local hospital. Charles Ornstein at ProPublica has some especially good ones.

    • Locate doctors or other trauma staff that may have been raising alarms. These people will often be named in lawsuits and whistleblower claims. Their contact information may also be contained city reports or legislative testimony.

    Other Datasets to Reference

    • The FBI’s Uniform Crime Reporting Program, Major Cities’ Chiefs’ Association, and local police compstat figures: These data sources will allow you to compare your figures against the annual counts that your police department is submitting for shootings in your area. You want to make sure that no incidents are missing from your analysis. The more sources you can use to check your numbers, the better.

    • Hospital intake data relating to gunshot victims: Analyzing hospital intake data for gunshot-related wounds can allow you to understand whether or not there are other factors leading to different mortality rates across the area you’re examining. You might find, for example, that people with pre-existing conditions are more likely to die from gunshots than those without. Generally, this kind of data is unavailable to journalists, but the closer to hospital data you can get, the more granular you can make your analysis, and draw stronger conclusions.

    • EMS incident response data: This data will allow your newsroom to cross-reference the case fatality rate with EMS response times. For example, if you find an area that has a higher gunshot fatality rate than other pockets of the city, and has less access to trauma care, this might result in higher EMS response times for that particular area. The data is available from local dispatch centers and is open to freedom of information requests. Some major cities also keep the datasets available online through open-data portals. The data should contain information for each EMS dispatch incident, with the key fields being: date and time the call was received, location of the incident, type of incident, time at dispatch, time on-scene, time leaving scene, time at hospital arrival, and incident status. For reference, NYC’s online EMS data contains many of the fields necessary to perform a competent response time analysis (link).

    • Gun Violence Archive: The Gun Violence Archive aggregates news reports about gunshot violence in databases that you can filter and interact with. If you read through and code each incident, you may be able to fill in holes in your data, or verify some of its findings. Consider reading through each news report for the information you are looking to check, for instance, which hospital was the victim taken to or how many times the victim was shot. You can then create a spreadsheet with fields such as “hospital taken” or “times shot.” This will enable you to understand which hospitals gunshot wound victims are being taken to, and determine whether there is a significant difference in where bullets are striking the victim. If there is an overwhelming amount of data and you have limited capacity to sift through it, consider examining the articles in the areas where there appears to be a higher mortality rate.

    Additional Story Threads to Consider

    • Have recent trauma center closures put communities at risk?

    • Are EMS response times particularly bad in a certain area?

    • Is there a hospital with significantly higher fatality rates compared to other hospitals?

    • Has the state been funding one area’s trauma care but not another?

    To get you started in your reporting, below are some research papers to read through and experts to contact. You are by no means limited to these sources or information, and we encourage you to look for more sources to build on your reporting.

    Relevant research

    Network Distance and Fatal Outcomes among Gunshot Wound Victims, August 2019 (Pre-print, awaiting peer review).

    Abstract: Despite nation-wide decreases in crime, urban gun violence remains a serious and pressing issue in many cities. Victim survival in these incidents is often contingent on the speed and quality of care provided. Increasingly, new research has identified the role that specialized trauma care plays in victim survival from firearm-related injuries. Using nearly four years of data on shooting victimizations in Philadelphia we test whether distance to the nearest level 1 trauma center is associated with victim survival. We employ different distance measures based on street network distances, drive-time estimates, and Euclidean distance - comparing the predictive accuracy of each. Our results find that victims who are shot farther from trauma centers have an increased likelihood of death, and drive time distances provide the most accurate predictions. We discuss the practical implications of this research as it applies to urban public health.

    Distance to trauma centres among gunshot wound victims: identifying trauma ‘deserts’ and ‘oases’ in Detroit, Injury Prevention, June 2019.

    Conclusions: Distance to the nearest trauma centre is associated with GSW victim survival. Clusters of block-groups with below-average GSW mortality were observed within close proximity of major trauma centres in Detroit. Improving speed and access to trauma care may play a role in reducing GSW mortality.

    Trauma Deserts: Distance From a Trauma Center, Transport Times, and Mortality From Gunshot Wounds in Chicago, American Journal of Public Health, 2013.

    Conclusions: Relative “trauma deserts” with decreased access to immediate care were found in certain areas of Chicago and adversely affected mortality from GSWs. These results may inform decisions about trauma systems planning and funding.

    Effects of closure of an urban level I trauma centre on adjacent hospitals and local injury mortality: a retrospective, observational study, BMJ Open 2016.

    Conclusions: During and after the MLK closure, trauma admissions increased at three of the four nearby hospitals, particularly admissions for gunshot wounds (GSWs). This redistribution of patient load was accompanied by a dramatic change in the payer mix for surrounding hospitals; one hospital’s share of uninsured more than tripled from 12.9% in 1999 to 44.6% by 2009. Overall trauma mortality did not significantly change, but GSW mortality steadily and significantly increased after the closure from 5.0% in 2007 to 7.5% in 2009.

    Rising Closures Of Hospital Trauma Centers Disproportionately Burden Vulnerable Populations, Health Affairs, 2011.

    Abstract: Closures of hospital trauma centers have accelerated since 2001. These closures may disproportionately affect disadvantaged communities. We evaluate how driving time between ZIP code areas and the nearest trauma centers—a proxy for access, given the time-sensitive nature of trauma care—changed nationwide during 2001–07. By 2007, sixty-nine million Americans (24 percent of the population) had to travel farther to the nearest trauma center than they did in 2001, and almost sixteen million people had to travel an additional thirty minutes or more. Communities with disproportionately high numbers of African American residents, uninsured people, and people living in poverty, as well as people living in rural areas, were more likely than others to be thus affected. Because mortality from traumatic injuries has also worsened for these vulnerable populations, policy makers should learn more about the possible connections—and consider such measures as paying trauma centers serving these communities higher amounts for treatment of injuries.

    Experts to contact

    Giovanni Circo is an assistant professor in criminal justice and forensic sciences at the University of New Haven. He received his Ph.D at Michigan State University. Circo conducted a study on gunshot fatality rates in Detroit and found that the further away someone was from a trauma center when they sustained a gunshot wound, the more likely they died. Circo is happy to help guide journalists through the analysis, and provide technical support if needed.

    Email: gcirco@newhaven.edu

    Andrew Wheeler (Ph.D) is an assistant professor of criminology and criminal justice at the University of Texas-Dallas. Wheeler worked with Circo on a study examining gunshot fatality rates and Philadelphia, which also found that the further away someone was from a trauma center when they sustained a gunshot wound, the more likely they died. Wheeler is also open to working with journalists to carry out this analysis, and pass along all the code he used for his research.

    Email: Andrew.Wheeler@utdallas.edu

    Feel free to reach out to us if you have any questions on developing this kind of a story, and good luck in your reporting!


    • Sean Campbell

      Sean Campbell is an investigative data journalist, adjunct assistant professor in the Columbia Graduate School of Journalism, and contributor to The Trace. His work has prompted action from members of Congress, change in the CDC, and contributed to policy changes at Twitter.

    • Laura Laderman

      Laura Laderman is a data analyst at Measure of America, a project of the Social Science Research Council. She performs quantitative analysis and creates data visualizations with a focus on mapping and geospatial applications. Follow her on Twitter @liladerm.

    • Maya Miller

      Maya Miller is now an engagement reporting fellow with ProPublica’s Local Reporting Network. When working on this story, she was a data and health journalism fellow with Measure of America, where she worked with newsrooms across New York City on data-driven health stories. Follow her on Twitter @mayatmiller.


    Current page