New York Trained a Million People to Reverse Overdoses. Its Own Records Counted 18,690

The state handed out 1,428,716 naloxone kits. It trained 1,030,738 people who were not police officers, paramedics or firefighters. Over the same eight years, the paperwork that follows a bystander overdose rescue came back 18,690 times.

Those three figures come from the same dataset, published Oct. 22 in Harm Reduction Journal by researchers at the New York State Department of Health's AIDS Institute, with Sharon Stancliff, the office's longtime medical director, as senior author. The study is a descriptive analysis of the state Opioid Overdose Prevention Program, which registers the organizations that give naloxone away and collects the forms that responders fill out afterward.

The arithmetic is not a scandal. The authors say plainly that reporting is voluntary and that the forms capture a fraction of what actually happens. But the gap between what the state distributes and what it can document is where the interesting material sits: it tells you which counties file paperwork, which settings generate overdoses, and which drug people are actually holding when they kneel next to someone on a floor.

What the Reporting System Measures

New York has been running this program since 2014, and it is a genuine piece of infrastructure. As of June 2025 it had more than 1,300 registered organizations operating nearly 6,000 sites, each with a clinical director on record, each recertified every two years. Registered programs receive naloxone free and hand it out with a short training; the standard allotment has been two doses per person. Under the state's standing orders, people with insurance can also get it at a pharmacy, and the Naloxone Co-payment Assistance Program covers up to $40 of the co-pay without any separate enrollment.

New York was early to this. A 2006 law let non-medical people administer naloxone during an overdose, the first wave of laws that eventually made lay reversal ordinary in most states. Two decades later the distribution curve is almost vertical. In 2016 the program moved 10,821 kits and trained 66,674 people. In 2023 it moved 440,930 kits and trained 243,264, growth the authors describe as more than forty-fold.

The pandemic bent the curve and did not break it. Between 2019 and 2020 kit distribution dropped about 20% and the number of people trained dropped by more than 40%. Both recovered through 2023 to the highest annual totals in the series.

The Gap Between Kits and Forms

After 2021 the number of kits distributed pulled away from the number of people trained, and the gap kept widening. The authors read that as a sign of secondary distribution: one trained person picking up supplies for friends, a spouse, a building's front desk. It is a reasonable interpretation and also an admission. Once a kit leaves the counter, the state has no record of who ends up with it, and the only evidence of use is a form that someone has to remember to return.

That is why the reversal count should be read as a floor. What it documents reliably is not how many overdoses New Yorkers reversed, but which parts of the system are disciplined about filling out paperwork in the days afterward.

---

The Overdose Moved Inside

The most useful number in the paper is where the reversal happened. Just over half of reported events, 53.5%, took place in a private residence. Another 21% happened in shelters and supportive housing, a share that rose sharply after 2018. Public locations, meaning parks, restrooms and businesses, accounted for about 17% combined. Service facilities such as syringe exchanges and treatment programs accounted for 1.6%, and those events presumably had trained responders standing nearby.

The supportive housing figure is the one that should reorganize how the state buys overdose prevention. Supportive housing is a health intervention in New York, funded through the settlement money and Medicaid, and it concentrates people with serious mental illness and substance use disorder under one roof with staff on shift. If one in five community reversals happens in that setting, then naloxone, training and post-overdose follow-up belong in the operating budget of every housing provider, not in a grant that has to be renewed.

There is a second signal buried in the same section. The share of reversal events in which someone called 911 rose from 48.6% in 2016 to 63.7% in 2023, with 54.7% across the full period. The authors credit Good Samaritan law education and greater trust in emergency response. The remainder still matters more: a reversal where nobody calls means a person wakes up in a stairwell with no assessment, no buprenorphine induction offered, no referral.

---

Who Is Being Revived Has Changed

The mean age of the person receiving community naloxone was 37.5, with a range from 4 to 90. Adults 25 to 34 made up the largest single group at 38.3%, followed by 35 to 44 at 25.1%. What moved over the eight years was the tail: the share of reversals involving people 45 to 54 and 55 to 64 rose steadily, and the share among 15-to-24-year-olds fell.

Males accounted for 70.3% of administrations and were older on average, 39.3 against 33.4 for females. The racial composition shifted more sharply than any other variable. White individuals fell from 80.1% of reported reversals in 2016 to 57.3% in 2023. Black individuals rose from 6.6% to 25.3% over the same period, and Hispanic individuals accounted for 18.3% overall.

The authors are careful here, and they are right to be. A rising share of reported reversals among Black New Yorkers can mean naloxone is reaching neighborhoods that were previously underserved, or that the overdose burden has moved, or that reporting organizations have changed. The paper's own framing points at the last one: the type of organization filing the form changed dramatically over the study period, and that reshuffles every demographic cut in the dataset.

---

Buffalo Fell Off the Map, and It Probably Did Not

Split the reports by region and something odd happens. New York City's share of reported reversals rose from 20.7% in 2016 to 45% in 2023, and ended up at 35.8% of all reports across the eight years. Syracuse climbed from 14.3% to 18.3%. Buffalo went the other way, from 23.3% of reports in 2016 to 5.8% in 2023. Albany, Rochester and Nassau-Suffolk drifted down more gently.

A naive reading is that western New York solved its overdose problem. The paper offers the correct reading instead, and it is about who files. Syringe exchange programs submitted 66.4% of every report the state received over eight years, and their share fell from 84.5% in 2017 to 47.9% in 2023 as housing providers, community organizations and other non-traditional partners joined the program. When the composition of reporters changes that much, the map changes with it. The authors say regional differences should be interpreted cautiously and may not correspond to real overdose incidence.

This is the structural weakness of voluntary reporting in one paragraph. A county can look like it is improving simply because its busiest reporting organization lost a staff member, and a county can look like it is deteriorating because a housing provider started filling out forms.

---

The Spray Won, and the Needle Still Matters

In 2016, intranasal naloxone at the 4 mg dose was under 1% of community administrations. By 2023 it was 74.1%. Intramuscular administration declined steadily, and older products such as the Evzio auto-injector and combination kits nearly vanished. Two doses were the most common amount given at 47.5% of events, with a single dose at 26.9%, which is the part of the data that speaks to potency: one spray is frequently not enough.

The authors argue against treating that shift as a straight upgrade. Intramuscular naloxone still has a case in an era of high-potency synthetic opioids, particularly for responders able to titrate a dose, and for people whose nasal passages are damaged from years of snorting or smoking. Their recommendation is to keep both routes available rather than let the nasal format crowd the other one out of the supply chain.

---

A Survival Rate That Hides a Drift

Among reversals with a known outcome, 95.5% of people survived, and the annual figure never left the 94.4% to 98.1% band. That is the strongest result in the paper and the easiest to misread. Community naloxone works, but the denominator is only the events where somebody was present, recognized what was happening, had a kit and reported it afterward.

Inside that flat line there is movement. Fatal outcomes were 1.7% of reports in 2016, rose to 4.3% in 2022 and settled at 3.5% in 2023, with 3% across the period and 1.5% of reports missing an outcome. The authors do not resolve this, and the honest reading is that two explanations are tangled: a more lethal drug supply, and a reporting base that expanded into housing and community settings where the person found down is more likely to be already in respiratory arrest.

---

What the National Numbers Say About Distribution

Four days before the New York paper appeared, a research letter in JAMA put a harder question to the whole strategy. Researchers from Virginia Commonwealth, Pittsburgh and Brown looked at 2.5 million nonfatal overdoses recorded by 5,324 EMS agencies between May 2020 and December 2025. After overdose deaths began falling in May 2023, nonfatal overdoses ran 20.9% below where earlier trends projected, opioid-involved overdoses 32.7% below, and naloxone administrations 21.3% below.

The striking figure is the last one adjusted for the others. Correcting for the fact that fewer overdoses means fewer chances to use naloxone, administrations held roughly steady, but use by bystanders was 28.6% lower than expected, even after accounting for the decline. Co-author Michael Barnett told Brown that a supply-driven decline in overdoses can also reverse.

Set the two papers next to each other and the conclusion is uncomfortable for everyone who spent a decade counting kits. New York's distribution machine is not the binding constraint. A million trained people and 1.4 million kits exist; the constraint is whether one of those people is in the room, or the hallway, or the shelter bunk two feet away, at the moment breathing stops.

What the State Can Actually Change

The authors' recommendations are modest and mostly cheap. Expand low-threshold access, which in practical terms means naloxone boxes and vending machines in places people already go at 2 a.m., an approach they describe as promising but not yet evidence-backed at scale. Strengthen peer distribution, because peers are the ones who reach the private residences that account for half of all reversals. And improve reporting itself, since a voluntary form returned days later is not a surveillance system.

For families in New York, the practical route has not changed. Naloxone is free through registered programs, available at pharmacies under standing orders, and the $40 co-pay assistance applies without enrollment. The state's program directory and request form sit at oasas.ny.gov.

What has changed is the argument for where the next dollar goes. The distribution network is built. The housing providers, the shelters and the peer networks are where the reversals are now, and those are the settings with the thinnest budgets and the least reliable paperwork.