In 2012, the Wall Street Journal described a “medical triumph.” The number of people shot and wounded had increased by nearly half over the previous decade, yet firearm fatalities had dropped precipitously, the news organization reported. The story cited data from the Centers for Disease Control, as well as an analysis from Howard University and Johns Hopkins University.

Among the factors cited as responsible for the trend: a proliferation of trauma centers, the adoption of bleeding-control techniques from the military, and a new approach to surgeries — more operations, spaced out over time, as opposed to one marathon session.

But the conclusion that advances in trauma care have dramatically improved the survivability of gunshot wounds is challenged by a new study published today in the American Journal of Public Health. The study, led by Philip Cook, a professor of public policy and economics at Duke University and a prominent gun-violence researcher, found that a firearm injury from an assault was just as likely to be fatal in 2012 as in 2003, and that more people were not, in fact, surviving gunshot wounds.

The study does not directly examine the role of trauma care in treating injuries, and its authors say it should not be interpreted as a critique or evaluation of medical practices.

A key flaw in the earlier research, Cook said, is that it inflates the number of people treated each year for nonfatal gunshot wounds in hospitals. That number did not jump nearly 50 percent, as previously thought — it stayed relatively flat, beginning and ending the 10-year stretch at about 41,000 cases each year.

“We were deeply misled by the available statistics,” Cook said.

Cook said he had been skeptical for years about the apparent upward trend of nonfatal firearm injuries. The only nationwide estimates of nonfatal firearm injuries come from the Centers for Disease Control and Prevention’s National Electronic Injury Surveillance System, or NEISS (“nice”) for short. NEISS draws on a nationally representative sample of hospital emergency departments, and it contains details about demographic characteristics, as well as the circumstances of the injury. The NEISS data indicated that the number of victims wounded in a gun assault had soared 49 percent from 2003 to 2012.

The substantial increase struck Cook as out of line with other sources of data on homicide, aggravated assault, and fatal injury. For example, in his paper he cites the American College of Surgeons’ National Trauma Data Bank, whose annual reports over the past 13 years reveal relatively steady case-fatality rates for firearm injury.

“What’s going on?” he wondered. “Why can [the jump in nonfatal injury] be so out of line with the other sources of information we have about gun violence?”

Motivated to take a closer look at the NEISS data, Cook and a research assistant, Ariadne Rivera-Aguirre, discovered two technical issues which they believe had distorted the numbers. The first was that an outsize number of cases had been coded as “unknown circumstances,” making it appear that there were more assaults in later years. The second issue was that, over the years, more than a dozen hospitals had dropped out of the sample. The hospitals that replaced them sometimes had many more firearm assaults than the hospitals that had dropped out, which also falsely inflated the data.

After adjusting for both issues and re-analyzing the data, Cook and his team found that the number of gunshot injuries had actually stayed relatively level.

Previous research conveyed the story that “there really is more gun violence, but the trauma care is more effective, so we have more survivors,” Cook said.

“In my study, my conclusion is, it’s really not that complicated,” he said. “There has been no increase in gun violence, and there’s also been no change in survival rates.”

From a public safety perspective, the finding that not as many people were shot in assaults as was thought should have be welcome news. But the implication that medical advances are not saving more lives than before is likely to be controversial.

Elliott Haut is a trauma surgeon at the Johns Hopkins Hospital in Baltimore. He reviewed the pre-published paper with Cook’s permission and said he does not agree with its conclusion. Haut said the research does not consider the severity of a gunshot wound. It doesn’t evaluate, for example, whether doctors are treating patients who have been shot more times, or who have been struck by bullets that cause more damage.

“The authors assume that the overall ‘severity’ of the gunshot wounds stays stable over time,” Haut wrote in an email. “All patients with gunshot wounds are not the same. They are not interchangeable.”

Haut recalled that when he was a medical student in Philadelphia in the mid-1990s, hospitals would routinely pump gunshot patients with liters of saline solution to help replace the blood they’d lost. But that procedure was found to disturb the clotting process, making patients bleed even more. Now, patients receive more blood products earlier.

The paper’s implication for the medical field, that changes in trauma care didn’t actually make all that much of a dent, strikes Haut as something that will be challenged by medical professionals.

“I think you would have a lot of very upset trauma surgeons, trauma centers, trauma nurses, who know we’re absolutely saving people we weren’t saving 10, 20 years ago,” he said.

Garen Wintemute, an emergency physician at the University of California-Davis, and co-author on the new study, said medicine has certainly made great strides in the past half-century, and that techniques continue to improve. But, he said, “it’s just not plausible” that those improvements “would account for a huge increase in survivorship from assaults.”

Though many studies have evaluated the effectiveness of individual medical interventions in treating gunshot wounds, Cook said he was not aware of any that measure the overall relationship between emergency or trauma medicine and firearm fatality. He thinks his findings send a clear message.

“If trauma physicians were out there taking credit for a big improvement in trauma care during this period, I think that they’ll stand corrected,” he said. “To have this circulating as a truth, that in fact there had been huge improvements for trauma care for gunshot cases, would be like claiming we had 100 percent effective treatment for the Zika virus.”