A few days after the San Bernardino mass shooting in December 2015, Dr. David Stark went for a run along San Francisco’s Embarcadero. He was listening to NPR when a segment came on about the lack of gun violence research funding from the U.S. Centers for Disease Control and Prevention.
Stark listened as the report told the story of the Dickey Amendment, the annual rider first inserted into the 1996 federal congressional appropriations bill prohibiting the use of CDC funds “to advocate or promote gun control.” Though not an outright ban, the measure has had a chilling effect on research.
At the time, Stark was a research fellow at Stanford University School of Medicine, studying biomedical informatics. He got back from his run, and soon got down to work on a side project.
“To me as a physician and as a research scientist, the idea that knowledge dissemination and knowledge generation was being systematically prevented, that was a pretty strong statement,” Stark recalled. “So I wanted to essentially test that hypothesis to see if that was born out in the data.”
Stark wanted to measure the effects of the congressional restrictions on gun violence research in statistical terms. To do that, he built statistical models to predict how much funding and how many published articles would be expected based on the number of people who died from 30 top causes of death. Data came from the Compressed Mortality File on a database known as CDC Wonder (Wide-ranging Online Data for Epidemiologic Research), the Federal RePORTER funding database, and the MEDLINE publications database.
Stark’s model assumed that the more Americans are killed by a given cause of death, the more the government will study that subject. Between 2004 and 2014, the United States saw about 350,000 deaths because of firearms, with a mortality rate of 10.4 deaths per 100,000 people. Based on how often people were dying from gunshots, Stark’s formula predicted nearly $1.4 billion in gun violence research funding and 38,897 publications.
In reality, gun violence research received only $22.1 million in federal funding and generated just 1,738 scientific articles during the decade in question. That shortfall became the centerpiece of widely covered analysis that Stark published last week in the the Journal of the American Medical Association.
During the period Stark studied, motor vehicle accidents claimed over 415,000 lives. That made for a death rate of 12.3 per 100,000 people — slightly higher than gun violence, which in recent years has caught up with car accidents as a killer of Americans — and predicted research funding of nearly $1.6 billion. Checking the actual research outlay, Stark found that motor vehicle accidents receive less robust funding than their death rates might suggest — but still much more than shootings.
The gap was even wider in knowledge generated on the two subjects: There were 44,710 separate publications on motor vehicle accidents in the literature, slightly more than the 43,390 predicted.
“I would hope that the data speak for themselves,” said Stark, 34, now assistant professor of health system design and global health at the Icahn School of Medicine at Mount Sinai Hospital in New York, where he also serves as medical director of the Institute for Next Generation Healthcare.
Motor vehicle accident deaths declined dramatically over the last three decades with progress made through policies like highway median dividers, drunken driving laws, emphasizing seat belt use and air bags. Stark believes that increased levels of research on guns could lead to a similar decline in mortality, while still respecting the Second Amendment.
“No one would say that the intent of traffic safety research was to eliminate automobiles from the road,” he said. “No one is necessarily saying that the intent of gun violence research is to eliminate guns. I recognize that that is a constitutional right. I actually grew up going to a summer camp on a riflery team. I was a pretty good marksman.”
Stark, who earned patches from the National Rifle Association as a child, says he had no political agenda as he worked intermittently on the study, his first work looking at gun violence.
But one cause of death included in Stark’s research, and highlighted in the study’s brief, angered critics.
“Gun violence killed about as many individuals as sepsis,” the study reads. “However, funding for gun violence research was about 0.7 percent of that for sepsis and publication volume about 4 percent.”
That comparison came from data in the Compressed Mortality File on CDC Wonder, which draws on information from all death certificates filed in the 50 states and the District of Columbia. But sepsis, a systemic inflammatory response to infection sometimes called blood poisoning, is difficult to diagnose, making tallies of deaths attributable to it imprecise.
“Sepsis could be due to a kidney stone, blunt force injury, a fall out a window or a stab wound,” said Dr. Brian L. Peterson, chief medical examiner for Milwaukee County, Wisconsin, who is president of the National Association of Medical Examiners. Clinicians sometimes disagree whether sepsis is even present in a given patient.
The CDC data used in the study recorded more than 391,000 sepsis deaths between 2004 and 2014, with a mortality rate of 11.6 deaths per 100,000 people. But a more recent CDC study in April 2016, using different data sources, estimated that sepsis deaths could actually be between 168,000 and 381,000 in a single year.
Stark acknowledged the higher estimates of deaths from sepsis when asked about questions his paper has generated. He said he used the sepsis figures from CDC Wonder because it is a robust data source allowing him to compare across many categories of death.
“For clarity’s sake, if I could write it all over again, perhaps I would emphasize motor vehicle accidents,” Stark said. “Because, number one, those figures are not at all contested.”