Mapping preventable death in “Bleeding Out”
By Lauren Caruba, The Dallas Morning News
In the back of an ambulance in San Antonio, I watched as paramedics worked on a man they had pulled from a house with bullet-riddled windows and blood-smeared tiles. He had been shot twice, in the arm and chest. When I looked down at my shoes, I saw the man’s blood spattered across my sneaker.
In that moment, I knew. The man was dying.
In most cities across the country, this patient’s treatment would have consisted of fluids, medications and breathing support. But these paramedics were equipped with the one thing the man really needed to survive, the crucial resource most ambulances lacked. Blood.
As the ambulance raced toward the hospital, paramedics grabbed a bag of blood from a cooler. The blood was still flowing, and the man was still alive, as he was rolled into the hospital.
That scene was the product of months of reporting on a program designed to deliver rapid transfusions to the patients who needed them the most throughout San Antonio and surrounding South Texas. But it was only the beginning of what would become a multi-year investigation of a national health crisis — an epidemic of preventable death in the American trauma system. Based on more than 140 interviews, the six-part series, “Bleeding Out,” was jointly published in late 2023 by The Dallas Morning News and the San Antonio Express-News.
I first learned about the blood program after attending a lecture at UT Health San Antonio, where local trauma surgeons spoke about the science of bleeding to death and the important role that rapid transfusions can play in saving lives. I spent months learning everything there was to know about how the program worked, interviewing trauma surgeons, blood banking professionals, trauma region coordinators, paramedics and patients. For more than 40 hours, photojournalist Jerry Lara and I shadowed San Antonio paramedics as they responded to shootings, stabbings, car crashes and other accidents.
Had my reporting only focused on this program, it would have yielded a strong explanatory story, but nothing more. Instead, I realized that if the trauma system in South Texas was doing something right, then the approach everywhere else was wrong.
Most investigations start with a problem, but this one started with the solution. I just needed to contextualize the larger problem and identify areas for accountability.
I needed to understand how and why these patients were dying, and what could be done to save them. Digging into the medical literature, I learned that physicians and paramedics who treated trauma victims had long discussed preventable bleeding deaths amongst themselves. Following a rabbit hole of footnotes, I eventually amassed almost 300 medical journal articles that formed the basis of my reporting.
I learned that there was a public health crisis hiding in plain sight.
Experts estimated that tens of thousands of injured Americans were bleeding to death with potentially survivable injuries, every single year. Every city and state was affected. In 2016, trauma experts testified before Congress about this unnecessary loss of life and the need for more funding for research, to little effect.
When someone is bleeding out, it triggers a cascade of breakdowns inside the body. With each minute that severe bleeding remains uncontrolled, the risk of death rises. When enough blood is lost, the body reaches a point of no return. The only way to reverse this process is to not only stop the bleeding, but replace what blood has been lost.
In trauma medicine, conventional wisdom used to dictate that patients had the best chances of survival if they were treated in the “golden hour” after sustaining a critical injury. However, more recent research indicated that the deaths of hemorrhaging patients peaked around the half hour mark. No matter how fast paramedics transported patients to the hospital, it wasn’t fast enough for these patients.
Yet across the country, the vast majority of ambulances did not carry blood products, despite military research showing that field transfusions can help buy valuable time for critically injured patients.
The authors of these research articles became some of my best sources. They had witnessed this phenomenon firsthand — as combat surgeons who transfused blood on the battlefield, as trauma surgeons whose patients got to them too late, and as paramedics who watched patients die because they didn’t have the resources to fully treat them.
The deaths of these patients were not inevitable. Our health care system simply was not properly equipped to save them.
I knew we needed an example of a preventable death. Using medical examiner data, I identified Dallas residents who had died in shootings and car crashes. After analyzing numerous autopsies, I narrowed my focus to Malik Tyler, a 13-year-old boy who had been shot and killed while walking home in 2019.
Trauma experts confirmed what I suspected — that his injuries from a single gunshot wound were survivable. If the paramedics had reached Malik faster and given him blood on the spot, the outcome could have been different.
What I still did not know was how regional disparities in trauma care helped drive this crisis.
During my research, I found a 2018 study linking poor trauma care access to more patients dying before ever reaching a hospital, based on death certificate data from CDC WONDER. These deaths were particularly prevalent in rural states, where there are fewer high-level trauma centers equipped to treat the most severely injured patients. Experts say higher rates of prehospital trauma deaths indicate an opportunity to save more lives, as many patients bleed out before they can make it to a hospital.
Enlisting the help of data journalist Ari Sen, I asked if the analysis could be updated and expanded upon. We reached out to the study’s author, University of Alabama at Birmingham assistant professor and trauma surgeon Dr. Zain Hashmi, who agreed to share his methods and review our analysis.
We wanted to illustrate how gaps in care contributed to worse outcomes for injured patients. We decided to map the nation’s trauma centers to identify trauma deserts, where bleeding patients would be at the highest risk of death. Building off Hashmi’s analysis, we would show that these care deserts corresponded with higher proportions of injured patients dying before ever reaching a hospital.
Our goal was to create a map that would show anyone in the United States roughly how far they were from a Level I or II trauma center, which provides the highest level of care to injured patients.
To start, we obtained a database of every trauma center in the U.S. through a licensing agreement with the American Trauma Society.
Sen initially tried using traditional geospatial methods like measuring the distance from the center of a Census Block to the trauma center, drawing circular buffers around each of the hospitals. But these traditional approaches had a critical flaw — they were measures of pure distance, not the time it would take for an injured patient to reach the hospital.
Eventually, Sen came across a blog post from Texas Christian University professor Kyle Walker on how to draw isochrones — shapes representing how far you could drive from a point in a given amount of time — with the Python programming language. Sen incorporated the code, drawing the shapes in 3-minute increments from 3 to 30 minutes away from every trauma center.
After applying some processing to make sure the isochrones for each hospital didn’t overlap, Sen faced another problem — what happens when someone is close to multiple trauma centers, as is often the case in big cities? With some help from a poster from an online forum, he developed a custom algorithm to deal with these intersecting areas.
For areas more than 30 minutes away from a trauma center, Sen downloaded a file with every Census block in the country and removed the Blocks that were covered by the isochrone shapes he had drawn. For the remaining Blocks, he computed the distance from the center of the block to the nearest trauma center.
Sen exported these shapes using a new file format called a geoparquet to significantly reduce the file size and loaded them onto a custom-built map. The map was embedded in the fourth story in the series, which focused on geographic disparities in trauma care and how where you live can determine whether you live after a major injury.
For that story, Sen also analyzed the data we created and used Census Block Group data to determine that roughly a quarter of all Americans lived outside of the critical 30-minute window to a high-level trauma center.
While Sen was conducting his analysis, photographer Smiley Pool and I traveled to western Colorado for a weeklong reporting trip, where we interviewed paramedics about the challenges of treating bleeding patients in extremely remote areas.
The remainder of the series focused on solutions, patient stories and barriers to improving patient care. Pool and I spent many hours embedded with frontline medical providers in North and Central Texas. We got to know a North Texas patient who wouldn’t have survived a head-on car crash without prehospital transfusions.
Before publication, Dallas Fire-Rescue committed to piloting a mobile blood program. Department leaders later estimated that prehospital blood would have benefitted nearly 700 patients in 2023 alone. This February, the program debuted, with blood available at two fire stations in North and South Dallas. At a press conference, city leaders said they expect the program will save lives.
That’s a sentiment that was consistently expressed to us as we investigated this problem — that our reporting not only could, but would help save lives. I can’t think of anything more rewarding than that.