Washington Told Treatment Programs What They Can't Say to Get Federal Money. The Deadline Was Yesterday.

Jack Martin runs Southside Harm Reduction Services out of a storefront in Minneapolis. Through July, Roll Call reported, federal dollars that used to help keep his lights on and his staff paid stopped being eligible to pay for either — a reversal of a 2016 congressional decision that first let federal money touch syringe service programs at all. Martin’s program is one of hundreds nationally that spent the past year learning to run on less, then on maybe, then on a legal theory about what “harm reduction” is allowed to mean in a federal grant application. The people who walk through his door are, overwhelmingly, people who are still using — showing up for clean syringes, wound care, a naloxone kit — not because they’ve decided to stop, but because staying alive has to come before that decision can happen at all.

On July 6, SAMHSA announced $281 million across fifteen discretionary grant programs — the largest single piece, $68.25 million, for medication-assisted treatment. HHS Secretary Robert F. Kennedy Jr. framed it as an investment in “treatment, recovery services, and tools to prevent overdose.” Applications closed yesterday, July 27.

Here is the sentence Filter found sitting in every one of those fifteen notices: the application and budget narrative must not support harm reduction.

Money that arrives with a gag order isn’t relief. It’s a filter — and it decides in advance which parts of the field get to survive.

That is the thesis, and it is worth sitting with before the numbers arrive, because the numbers alone can read as abstract. A “harm reduction” prohibition sounds like a values dispute, the kind of Washington argument that stays in Washington. It is not. It is an instruction to the people writing these applications — the case manager drafting the narrative, the program director translating what her outreach team actually does into language a federal reviewer will accept — to describe their work in a way that omits the parts most likely to keep someone alive tonight. Naloxone distribution. Fentanyl test strips. Wound care for people who are still using. Say it plainly, in the language people in recovery actually use with each other: you can’t write “we help people not die while they’re still using” on a form the government will pay you for. You have to write around it, or you don’t get paid.

The plain version, first: what “harm reduction” means on this form

Before going further, the translation the funding notices skip. Harm reduction is not a philosophy about whether drug use is acceptable. It is a set of practices — sterile syringes, naloxone, fentanyl test strips, wound care, nonjudgmental contact — built on a specific, boring, well-evidenced premise: a person has to be alive to recover, and most of the deaths happen before anyone gets the chance to offer treatment. Public health researchers didn’t invent this to make a political point. They invented it because syringe exchanges measurably cut HIV and hepatitis C transmission, because naloxone reverses overdoses in minutes, and because both interventions work on people who are not, yet, ready or able to stop using. That’s the entire argument. What SAMHSA’s fifteen notices now say is: describe that work without describing it.

We have run this experiment before, and we know how it ends

This is not the first time the federal government has decided that a cheap thing which keeps people alive sends the wrong message. In 1988, at the height of the AIDS crisis, Congress banned the use of federal funds for syringe exchange — then one of the only interventions slowing HIV transmission among people who inject drugs. The ban held for 21 years, through Republican and Democratic administrations alike, lifted only in 2009 and then reinstated in part before Congress finally let federal dollars flow to exchange operations (though never to syringes themselves) in 2016 — the same 2016 decision that just got partially reversed. Public health researchers spent those 21 years counting HIV infections that modeling consistently showed a funded, unrestricted exchange system would have prevented. Nobody who lived through that fight disputes, in retrospect, that the ban cost lives to make a point. The 2026 version of the argument has different packaging — “the budget narrative must not support harm reduction” is bureaucratic where 1988’s ban was legislative — but it is the same bet: that withholding the tool sends a clearer message than the deaths it produces.

Public health researchers spent those 21 years counting HIV infections that modeling consistently showed a funded, unrestricted exchange system would have prevented.

The number attached to the bet

For the first time, that bet has a specific, peer-reviewed price tag. A modeling study published in JAMA Network Open in June — led by researchers including Czarina Fetters, tracking a simulated cohort of 3,694,500 people who inject drugs from August 2025 through August 2030 — projected the health effects of the funding disruption that followed last year’s executive order halting federal support for syringe services. Under the study’s worst-case scenario, sustained high service disruption over five years, the model projected 39,600 additional deaths, including 15,600 additional overdose deaths, with all-cause mortality up to 5.0% higher and overdose mortality up to 6.9% higher than it would otherwise be. Hepatitis C, meanwhile — a disease that still kills more than 11,000 Americans a year despite having a cure, according to Roll Call’s reporting on the CDC data — is projected to climb as the syringe access that cuts transmission by roughly half becomes harder to reach.

Regina LaBelle, a Georgetown addiction-policy researcher who served as acting director of the Office of National Drug Control Policy under President Biden, was blunt in that same reporting: this isn’t a hypothetical fight over language, it is a countable one. Sen. Edward Markey led a group of senators in a letter warning the administration that the policy “puts American lives at risk” — not as a rhetorical flourish, but as a description of a number that now exists in a peer-reviewed journal.

The frustrating part, for anyone who has spent time in this field, is that the evidence on the other side of the ledger has never been stronger. A prospective cohort study just published in Addiction — the NalPORS study, following more than a thousand people who use opioids across 22 harm-reduction and treatment sites in England, Wales, Scotland, and Sweden — found that when people who had received take-home naloxone actually witnessed an overdose, 83% of the time naloxone was administered, safely, by a peer with no medical training. That is not a marginal statistic. It is close to as good as any medical intervention gets, delivered by exactly the population SAMHSA’s new grant language makes it harder to fund. Overdose deaths nationally are, encouragingly, falling — provisional CDC data released in June put the 12-month total at 69,147 deaths, a 13.2% year-over-year drop. The honest reading of that decline credits several forces at once — supply disruption, wider naloxone access, better treatment access — and cutting off one of the tools that’s working, in the middle of a genuine improvement, is not a neutral move. It’s a bet that the trend will hold without the piece being removed.

What programs are actually doing about the sentence

Filter’s reporting traces this back to an April 2026 guidance update tightening the language of last year’s “Ending Crime and Disorder on America’s Streets” executive order, and notes that SAMHSA separately pulled fentanyl test strips from funded-activity eligibility that same month. Programs that spent the past decade building harm reduction into their core model — not as an add-on, but as the front door that gets someone to treatment later — are now writing grant narratives that describe outreach vans, wound care, and naloxone distribution using language stripped of the words a reviewer might flag. Executive directors are asking staff to functionally translate their own work into euphemism to stay fundable, while the actual practice on the ground, per every program reached in Filter’s and Roll Call’s reporting, hasn’t changed. It cannot change — the same people are still walking through the door, still using, still needing supplies that keep them alive long enough to consider anything else. What has changed is whether the federal government will admit, on paper, that this is happening.

It cannot change — the same people are still walking through the door, still using, still needing supplies that keep them alive long enough to consider anything else.

There’s a real internal tension worth naming here too: RFK Jr. himself publicly endorsed methadone and buprenorphine at last year’s Rx Summit — the two medications STAT reported are associated with roughly a 50% lower overdose death risk — even as SAMHSA’s April “Dear Colleague” letter cautioned providers against their long-term use, and even as this same $281 million cycle asks applicants to write around the harm-reduction practices that, per the evidence above, function as the bridge to those medications for people not yet in treatment. An administration can’t credibly claim to champion medication for opioid use disorder while defunding the outreach that gets people to a prescriber in the first place.

If you are reading this in recovery, or close to someone who is: the things that are still yours haven’t disappeared. Naloxone remains legal to carry and, in most states, available over the counter without a prescription. Test strips remain legal to possess almost everywhere, even where federal dollars can no longer pay to hand them out — community programs are finding other funding to keep them stocked. 988 is still answered. None of the things this piece just described take away your ability to carry naloxone tonight, or to reach someone who will help you use it.

What this fight is actually about is who pays for the next round, and whether the people doing that work can say, on the record, what they’re doing it for. In 1988, the ban held 21 years and researchers counted the infections it caused the whole time. The 2026 version isn’t a ban — it’s a sentence embedded in fifteen grant notices, technical enough that most of the country will never read it, that asks people like Jack Martin to keep doing the same work while describing it as something else. He’ll keep doing it. The question this July 27 deadline actually answers is how many programs like his can afford to.