JAMA Study Finds Naloxone Use Did Not Rise as Overdose Deaths Fell

Naloxone is not why American overdose deaths fell. That is the conclusion of a JAMA research letter published Oct. 21 that tested the most widely repeated explanation for the steepest decline in drug mortality in three decades and found the medication being used at roughly the same rate per overdose as before the drop began.

Researchers from Brown University's School of Public Health, Virginia Commonwealth University and the University of Pittsburgh examined records from 5,324 emergency medical services agencies covering May 2020 through December 2025, a window that brackets the turn in overdose deaths in May 2023. The dataset included 2.5 million responses for nonfatal overdoses.

Comparing the period before and after that turning point, nonfatal drug overdoses ran 20.9 percent below what earlier trends projected. Opioid-involved overdoses ran 32.7 percent below. Naloxone administrations did not rise to fill the gap: they fell 21.3 percent below projected levels, which is close to what a smaller number of overdoses would produce on its own.

The test the naive comparison misses

A falling count of naloxone doses is not evidence about naloxone's usefulness, because fewer overdoses mechanically mean fewer doses. The researchers addressed that by measuring administrations per overdose rather than in total. On that basis the medication was reaching people at nearly the same rate as before the decline.

The one component that moved the wrong way was bystander use. Doses given by friends and family members ran 28.6 percent below what earlier trends predicted, even after accounting for the smaller number of overdoses.

"There are big questions about why opioid overdose mortality rates are falling nationally since 2023," said Dr. Michael Barnett, a health services researcher at Brown's School of Public Health and a physician at Brigham and Women's Hospital, in the university's announcement. "A lot of people think that increased use of the overdose reversal agent naloxone is a big reason. This study changes the narrative and says we have to look to other causes for overdose declines."

What the data cannot see

The study's own limits are unusually consequential for how the finding should be read. The authors state plainly that they did not examine the competing explanations — changes in the illicit drug supply, or changes in how and by whom opioids are used — and therefore have not identified what did drive the decline.

The records also exclude the overdoses that never reach the emergency system, including reversals performed by bystanders who did not call 911. Since bystander naloxone has been the centerpiece of state distribution programs, that blind spot is exactly where a naloxone effect would most plausibly appear. A rise in neighborhood-level reversals handled without an ambulance would not show up as a rise in EMS administrations.

Why the mechanism matters more than the arithmetic

If naloxone had driven the decline, the improvement would be something public health agencies could defend and extend by buying more of it. If the driver is drug supply — a shift in what is sold as fentanyl, its potency, or its availability — then the decline is a property of the criminal market rather than of public policy, and it can end without warning.

Barnett's own reading points that way. "My own read is that a disruption in the fentanyl supply best explains how sudden and widespread the turn was," he told Science News, adding that the population at risk "has probably been shrinking for years." He also warned that "a supply-driven decline can reverse," with traffickers able to adapt by finding new chemical sources or turning to more potent synthetic opioids. The national decline did slow in 2025, which is consistent with that caution.

That framing does not diminish naloxone. It separates two different claims that have been fused in public messaging: that naloxone saves individual lives, which remains true and is not in dispute, and that naloxone distribution explains a population-level mortality trend, which this data does not support.

The policy consequence

For agencies that spent the past three years scaling naloxone distribution, the finding argues for keeping the investment while reopening the question of what else is working. National deaths fell from roughly 111,000 a year at the mid-2023 peak to under 71,000 by the end of 2025, a decline large enough that its cause carries real weight in budget decisions.

The study also gives harm reduction a harder measurement problem. If the number of overdoses is falling because the supply is changing, then surveillance of the drug supply — what is in the bag, how potent it is, what adulterants are circulating — becomes as important as the count of reversals. Several states have moved in that direction, and the national picture of a declining fentanyl toll now takes on a different explanation than the one most often offered.

The letter's authors close by calling for research into what is driving the decline and how public health programs can sustain it. Until that work exists, the honest summary of the 2023 turn is that it remains unexplained — and that the explanation most people assumed has now been tested and did not hold.