The Overdose Numbers Are Finally Falling. Washington Just Defunded the Reason Why.

Nabarun Dasgupta met Tony Giveds at Yale in 2002. Giveds worked outreach — the guy who actually talked to people using drugs on the street, not the guy who studied them from a lab. He was the first person who really connected Dasgupta with the human side of the drug problem in the United States. In 2004, after a bad stretch, Giveds died by intentional overdose. “It was just super hard to feel him disappear from my life,” Dasgupta has said. He turned the grief into data work — years of it — and now runs the University of North Carolina’s Opioid Data Lab, tracking what’s actually killing people and what’s actually keeping them alive.

The number Dasgupta’s field just posted is real. So is the reason it’s falling. And this spring, the federal government defunded the reason.

This summer, for the first time in his career, Dasgupta got to call something a “moment of hope.” Provisional CDC data released in June counted roughly 69,147 overdose deaths for the twelve months ending January 2026 — a 13.2% drop from the year before, and the third straight year of decline from the 2022 peak of nearly 108,000. A more recent provisional cut, for the period ending in February, shows the decline holding at around 12%. Brandon Marshall, a Brown University researcher who has tracked overdose trends for a decade, put it carefully: “I’m cautiously optimistic that this represents really a fundamental change in the arc of the overdose crisis.” He also named the catch, in the same breath: “If deaths are going down rapidly, that means they can increase just as rapidly if we take our foot off the gas.” Kimberly Douglas knows both halves of that sentence from the inside — she founded Black Moms Against Overdose after losing her 17-year-old son, and she credits the harm-reduction infrastructure specifically: “We are starting to see overdoses go down in some places and that’s because of harm reduction.”

The forensic data backs the hopeful read up to a point. Naloxone is available over the counter. Settlement money, treatment expansion, syringe services, and drug-checking programs are, together, doing the unglamorous work of not letting people die — and the drop spans fentanyl, cocaine, and methamphetamine deaths alike, not just one drug class getting lucky. But Alex Krotulski, who directs the Center for Forensic Science Research and Education and watches what’s actually showing up in the drug supply, is blunter about how fragile the trend line is: “The drug supply continues to change and evolve.” His lab identified 23 new drugs in the first five months of 2026 alone — more than the 27 it found in all of 2025 combined — including a synthetic opioid called cychlorphine, reportedly up to ten times stronger than fentanyl and already surfacing as an undisclosed cutting agent. And the national number is not evenly distributed: seven states saw overdose deaths rise, several by more than 10%, and three of them — Arizona, Colorado, and New Mexico — are states where Rize Recovery works or plans to launch. A national headline reading “down 13%” can describe a country where Kentucky is safer and Arizona is not. Both things are true in the same dataset.

That’s the sentence to sit with, because it is the whole story. On April 24, 2026, SAMHSA sent grantees a Dear Colleague letter announcing that federal dollars could no longer pay for fentanyl, xylazine, or medetomidine test strips, syringes, or other supplies the agency now describes as “harm reduction… practices that facilitate illicit drug use.” The letter frames the shift as consistent with an executive order on ending disorder in American streets, and as a return to “evidence-based” priorities — a strange choice of word, since fentanyl test strips have a substantial evidence base behind them and SAMHSA itself was funding them as recently as July 2025.

That’s the sentence to sit with, because it is the whole story.

What that letter actually did, in one Kentucky county

Shreeta Waldon runs the Kentucky Harm Reduction Coalition. On April 25 — the day after the letter went out — she learned her organization was losing a $400,000 SAMHSA grant. In the first quarter of this fiscal year alone, the coalition had distributed 48,465 fentanyl test strips to people who wanted to know, before they used, whether what they’d bought was going to kill them. At the time the grant was pulled, Waldon estimated the coalition had about a month of test-strip supply left. “It doesn’t make sense that one day something is an evidence-based protocol,” she told CBS News, “and you decide, because of political climate, it is no longer evidence-based.”

Waldon’s organization was not alone and not the smallest hit. A South Carolina outfit called Fyrebird Recovery lost a $4,000 grant — a rounding error in the federal budget, the whole operating margin for a group that size. Its founder, A’zhane Powell, put the fear plainly: “How far will it go until we’re back to square one again?” At the Drug Policy Alliance, federal policy director Maritza Perez Medina described the reaction among people who’d spent a decade building this infrastructure: “People are just astonished.” She added the part that should worry anyone tracking the CDC numbers: “Overdose deaths have come down from the peak that we saw during COVID, and that’s great” — the implicit rest of the sentence being and this is how you put that at risk.

If you’ve ever used a test strip before you dipped in — dissolved a little in water, waited the two minutes, watched for one line or two — you already know what a few dollars of paper buys: the difference between a decision and a coin flip. That’s not abstract policy language. That’s the actual transaction SAMHSA’s letter interrupted, multiplied across every Kentucky Harm Reduction Coalition and Fyrebird Recovery in the country.

We have done this before, and we counted the bodies

This is not a new argument in the fight against overdose death, which is what makes it worth the six paragraphs of history. In 1988, at the height of the AIDS crisis, Congress banned the use of federal funds for syringe exchange programs — the intervention public health officials were, at that exact moment, identifying as the single most effective tool against HIV transmission among people who inject drugs. The ban held for 21 years, lifted only in 2009, after a generation of epidemiological data had accumulated showing what the delay cost. Researchers didn’t get to run a controlled experiment on how many HIV infections that policy caused; they only got to count, retrospectively, the gap between what syringe access could have prevented and what it was allowed to prevent. The pattern — a cheap, evidence-backed tool for keeping people alive gets recast as “enabling,” and gets cut, and the cutting looks decisive right up until the data catches up to it — is not new. It is, historically, the default response to progress against an epidemic among people society has decided are not fully sympathetic. Fentanyl test strips are the 2026 version of the 1988 syringe.

For the people running programs right now, the move is triage, not despair

If you’re a case manager or a program director reading this over coffee, the practical question is not whether the politics are right — it’s what fills the gap this quarter. A few things are worth raising at your next team meeting. First: verify what state or opioid-settlement dollars in your jurisdiction can legally substitute for the lost federal line — Arizona’s county-level settlement dashboards, for instance, are quarterly and public, and several counties have discretionary categories that test strips and safer-use supplies can fit under even after this SAMHSA shift. Second: naloxone itself is a different funding stream and, as of this writing, is not touched by the April letter — if your organization has been bundling naloxone distribution and test-strip distribution under one grant line, it’s worth confirming which piece survives before you assume the whole program is dark. Third: document the gap. Waldon’s “one month of supply” estimate became a citable, quotable fact precisely because she’d been tracking distribution numbers closely enough to know it. If your program loses a funding line, the number of people you can no longer reach is data somebody — a funder, a reporter, a state legislator — will eventually want, and having it ready is worth ten minutes now.

Waldon’s “one month of supply” estimate became a citable, quotable fact precisely because she’d been tracking distribution numbers closely enough to know it.

None of that fixes the policy. It just means the four weeks between a funding cut and a program actually going dark don’t have to be wasted.

The part that’s still yours

Here’s what the April letter did not touch, and what stays true regardless of who’s in charge of SAMHSA’s messaging this quarter: naloxone is still federally supported and still free or low-cost in most states, through channels separate from the harm-reduction supply cuts. If you carry it, you can still get more. If you don’t carry it and you’re around anyone who uses opioids, this is still the week to start. That’s not a caveat tacked onto the end of a policy story — it’s the one lever in this whole mess that a person can still pull today, no grant renewal required.

Dasgupta’s data lab is still counting. So is Waldon’s coalition, one dwindling box of test strips at a time, and so is Fyrebird Recovery, and so is every program that just found out the “evidence-based” label is retractable depending on who’s reading the evidence. Tony Giveds has been gone for twenty-two years, and the numbers Dasgupta built in his memory just posted their best year since anyone started counting this way. Whether next year’s numbers still look like a moment of hope is now, mostly, a question of whether the people who cut Kentucky’s grant understand what they actually cut — and whether anyone makes them look at the count in twelve months, the way Dasgupta has spent two decades teaching people to.