Overdose Deaths Are Falling Fast, and It Wasn’t the Crackdown That Did It

Tripti Choudhury spends her outreach shifts in Phoenix parks that are, increasingly, off-limits to her. She’s a trauma clinician and a volunteer with Shot in the Dark, a harm reduction group that meets people who use drugs where they are — a park bench, a wash, a sidewalk — with naloxone (the overdose-reversal medication sold as Narcan), sterile syringes, and test strips that reveal whether a supply is cut with xylazine, an animal tranquilizer naloxone can’t reverse. New Phoenix park ordinances now restrict where she can even work. “We’re restricting where we can go,” she told KJZZ in March. “When they’re on their feet… we can’t serve them.”

That’s Arizona right now: overdose deaths up 18% in a year, one of a handful of states moving the wrong direction while the country posts its steepest overdose decline in a generation. Nationally, provisional CDC data count roughly 67,798 deaths in the 12 months ending March 2026, down 12.3% from a year earlier — after a 27% single-year drop in 2024 that the Commonwealth Fund called sharper than three decades of prior data.

This decline was built by naloxone, addiction medicine, and quiet prison reform — not by the people now trying to take credit for it.

It’s also a decline with a warning built into it: the people who did the work are being defunded, state by state and grant by grant, and Arizona previews what the rest of the country risks looking like if that keeps happening.

The Drop Has a Cause List, Not a Hero

Ask the researchers who study this and you get the same answer in different words: no one thing did this. Evan Gumas, a senior research associate at the Commonwealth Fund who co-authored the September brief, put it to STAT News plainly: “There is no one thing that led to the decrease.” Brandon Marshall, an epidemiologist at Brown University, called the speed of it something he hadn’t seen before: “It is unprecedented to see overdose deaths come down so dramatically, so quickly.”

The list Gumas and his colleagues assembled is long, but four items on it do most of the work. First, naloxone saturation — the drug is now stocked in vending machines, carried by police and firefighters, and handed out by outreach workers like Choudhury in volumes that didn’t exist five years ago. Second, wider access to methadone and buprenorphine, the two medications for opioid use disorder (MOUD) that reduce overdose risk by curbing withdrawal and cravings without producing the same high as the drug someone was using; Colorado alone doubled its number of treatment clinics offering these medications, from 25 to 54. Third, treatment inside jails and prisons, where people are traditionally cut off from medication the moment they’re booked — New York saw a sevenfold increase in incarcerated people receiving MOUD between 2022 and 2025. Fourth, better data: state health departments can now track overdose deaths and near-misses fast enough to redirect naloxone and outreach funding to where a bad batch of drugs just landed, instead of finding out months later.

Medicaid sits underneath all four. Gumas called it “the single most influential lever for coverage and access” — the program that actually pays for the buprenorphine prescription, the outreach worker’s salary, the ambulance ride. That’s also why the same brief is not purely celebratory. Sarah Wakeman, an addiction medicine physician at Mass General Brigham, told STAT that “every single overdose death is really a failure, because no one should die from an opioid-related overdose when we have an antidote” — a reminder that roughly 68,000 people still died, even in the best year this crisis has had. And a Penn/Boston University estimate cited in the same reporting found that pending Medicaid cuts could strip treatment access from 156,000 people who currently have it, potentially doubling the overdose rate in that group alone.

If you’re one of the people this data is actually about — on methadone, on buprenorphine, carrying naloxone in your bag because someone taught you how — none of this is abstract. It’s the reason you’re statistically safer this year than you would have been five years ago, and it’s also the reason that safety isn’t guaranteed to hold.

It’s the reason you’re statistically safer this year than you would have been five years ago, and it’s also the reason that safety isn’t guaranteed to hold.

Washington Tried This Before, and It Cost Lives It Didn’t Have To

There’s a precedent for what’s happening to fentanyl test strips right now, and it’s not flattering. In 1988, at the height of the AIDS epidemic, Congress — led by Senator Jesse Helms — banned the use of federal funds for needle and syringe exchange programs, on the theory that providing sterile syringes to people who inject drugs would encourage more drug use. The evidence said otherwise almost immediately: syringe programs reduced HIV and hepatitis C transmission without increasing drug use, a finding that piled up for two decades while the ban stayed in place anyway. Congress didn’t lift it until 2009, and even then only briefly — it was reinstated in 2011 amid the same objections that created it in the first place, according to CSIS’s history of the policy. People got infected with diseases that a piece of federally funded plastic could have prevented, for a reason that had nothing to do with evidence.

On April 24, 2026, SAMHSA sent grantees two “Dear Colleague” letters — the formal notices federal agencies use to tell grant recipients what they will and won’t fund going forward — announcing what STAT described as a “clear shift away from harm reduction and practices that facilitate illicit drug use.” The letters banned federal money for fentanyl, xylazine, and medetomidine test strips, along with sterile syringes, safer smoking supplies, and overdose hotlines that keep someone on the phone while they use so help can be sent if they go quiet. Naloxone distribution was explicitly preserved. So was professional-use testing for law enforcement and medical staff — just not the version a person uses on their own supply before they use it. Health Affairs Forefront documented how the same guidance also discourages treating MOUD as anything but a bridge to abstinence, describing it as something that should not become “a default sentence to life-long medication use” — language that runs against the clinical consensus the addiction medicine field has spent years building.

The parallel isn’t exact — fentanyl test strips aren’t syringes, and 2026 isn’t 1988. But the shape of the argument is the same: a low-cost tool that keeps people alive long enough to get better, cut off because officials believe, without the data behind them, that it sends the wrong message. Test strips work the same way syringe exchanges did — they don’t make anyone start using drugs, they just mean the person already using finds out what’s actually in their supply before it kills them.

Arizona Is the Preview

Arizona recorded 2,531 overdose deaths in 2024 and 2,988 in 2025 — an 18% increase, according to the Arizona Public Health Association, driven by a nearly 33% jump in synthetic-opioid deaths and a roughly 70% jump in cocaine-involved deaths. New Mexico and North Dakota are the only other states moving the same direction; nearly everywhere else, states like Alabama, New York, and Virginia posted 25–30% drops, NPR reported in May, citing New York Governor Kathy Hochul’s claim that opioid deaths there have been “cut in half since 2022.”

Beth Meyerson, who directs the Harm Reduction Research Lab at the University of Arizona’s College of Medicine, told NPR the state is fighting the trend with fewer of the tools that are working elsewhere: fentanyl and methamphetamine circulating in more potent forms, rural opioid treatment that can require a two-hour drive, and overdose mortality among Native Americans that runs double the national rate. National Institute on Drug Abuse director Nora Volkow, in the same reporting, called methamphetamine’s role in Arizona and New Mexico “very problematic” — a second drug supply crisis riding alongside fentanyl that naloxone alone can’t fix, because naloxone only reverses opioids.

Dr. Matt Evans, medical director of addiction medicine for Circle the City in Phoenix, gave KJZZ the plainest version of why medication access matters so much: people on methadone are “60% less likely to have an overdose,” and people on buprenorphine are “40-plus percent less likely,” compared with people who aren’t on either. Arizona has both medications available — what it doesn’t reliably have is the funding, staffing, and outreach access to get them to the people who need them, especially as displacement from encampments and park ordinances scatters the population outreach workers are trying to track. That’s the mechanism connecting Choudhury’s shrinking map of where she’s allowed to work to the state’s rising death count: harm reduction doesn’t fail because the tools stop working, it fails when the tools can’t reach anyone.

Who’s Actually Taking the Credit

The politics of this moment are worth naming plainly. The same administration whose April 2026 SAMHSA letters cut funding for fentanyl test strips also frames its approach to the overdose crisis as a law-and-order win — tougher enforcement, tougher borders, a “clear shift away from harm reduction,” in the agency’s own words to STAT. Acting SAMHSA leader Chris Carroll signed both letters. SAMHSA cancelled roughly $1.7 billion in block grant funding, cut $350 million specifically from overdose prevention programs, and cut its own staff by more than half in the same stretch, per STAT’s April reporting. None of that shows up in the causal list researchers actually point to. Naloxone saturation, MOUD expansion, prison treatment reform, and Medicaid coverage do — and every one of those is a program built and funded by the public health infrastructure the same letters are dismantling.

None of that shows up in the causal list researchers actually point to.

That’s the uncomfortable arithmetic underneath the good news: the decline is real, but it was produced by exactly the interventions currently losing their federal funding. NIDA director Nora Volkow told NPR the drop only recently looked durable enough to trust: “We started to see the declines in 2023, but they were small and we weren’t certain they were going to be sustainable.” Nobody in that research community is crediting a crackdown. Crediting one anyway for a drop that harm reduction workers, addiction physicians, and Medicaid built is not just inaccurate — it hands political cover to the exact cuts most likely to reverse it, and it erases the people, like Choudhury, who actually did the work in the parks and clinics where it happened.

Naloxone is still legal and still free in most states, sitting on pharmacy shelves without a prescription in Arizona and nearly everywhere else, tonight, regardless of what happens to test strip funding. If you’re on methadone or buprenorphine, that medication is still doing exactly what Dr. Evans described — cutting your overdose risk by 40 to 60% — and no federal letter has touched your access to it. The tools that built this decline haven’t all disappeared. Some of them are simply harder to reach, in Arizona and in the parks Tripti Choudhury isn’t allowed to work anymore, than they were a year ago.

The country reversed the biggest drug death crisis in its history without a single dramatic law getting passed — it did it with a nasal spray, a clinic appointment, a syringe program, and someone willing to show up. It can un-reverse the same way, one closed access point at a time, if the people who did the actual work keep getting priced out of doing it.