A New Study Finally Measured Why Naloxone Access Actually Prevents the Next Overdose

Everyone in harm reduction has believed for years that getting naloxone into more hands prevents more deaths. Belief and measurement are different things, and a study published this spring in the Harm Reduction Journal is one of the clearer attempts yet to actually measure it — using the kind of data that can tell the difference between “these two things happen together” and “one of them is causing the other.”

Getting naloxone at a syringe services program doesn’t just correlate with fewer future overdoses. In this cohort, it measurably predicted them.

What the researchers actually did

A team led by Marina Plesons, with Tyler Bartholomew, Hansel Tookes, and Daniel Feaster, followed 136 clients of IDEA Miami, a syringe services program, using eight years of the program’s own administrative data (2017–2025). Their method was a cross-lagged panel model — built specifically to answer the causality question rather than dodge it, by checking whether Variable A at one time point predicts Variable B at the next time point, after accounting for how much each variable predicts itself over time. That structure is what lets researchers say “naloxone receipt came before the drop in overdoses,” not just “the two showed up in the same dataset.”

The results held up across three sequential quarterly assessments. Recent overdose experience strongly predicted receiving naloxone at the following visit (standardized estimates of 0.21 to 0.41, p<0.001) — programs are, unsurprisingly, good at getting naloxone to the people who just needed it most. But the arrow ran the other way too: receiving naloxone at an SSP visit predicted a lower chance of reporting an overdose at the next assessment (estimates of −0.18 to −0.28, p between 0.002 and 0.004). Descriptively, the pattern is stark — the share of participants reporting a recent overdose fell across the three waves from 43% to 31% to 28%, while the share who had naloxone on hand when they needed it climbed from 53% to 69% to 79%. Naloxone receipt itself stayed roughly flat, near 51%, meaning the program didn’t need to hand out more of it to see the effect — it needed the same supply to reach the same people consistently.

What it doesn’t prove, and why that matters more than usual

The authors are direct about the limits, and worth taking as seriously as the finding itself: 136 people at a single Miami program is not a national sample, and the effect size, while statistically real, comes from administrative program data with confounders the design can’t fully rule out — a pandemic and a rapidly shifting drug supply both sit in the background of 2017–2025 unmeasured. Self-report carries its own bias. And structurally, the study can only see people who survived to their next assessment — anyone who died of an overdose between visits drops out of the data, which means the “protective effect” measured here is likely a floor, not a ceiling, on naloxone’s real impact.

That caveat lands differently this year than it might have in a quieter one. U.S. overdose deaths fell 27% in 2024, the sharpest national decline in a generation, and CDC’s own provisional data confirm the scale of the drop. No single study explains a national trend, and Brown University’s Brandon Marshall was right to caution that the leading hypothesis points mostly to changes in the drug supply itself. But the Plesons study is one of the few pieces of evidence that connects the national number to an actual mechanism at the level of a single person’s next visit to a single program — which is a different, harder kind of evidence than a correlation between “more funding” and “fewer deaths” nationally.

It also matters what the people actually receiving naloxone say about it. In a separate qualitative study published this year in the journal Addiction, researchers interviewed 22 people who use drugs in Maryland, all of whom had personally reversed someone’s overdose with naloxone in the past three months. Their skepticism wasn’t about whether to carry it — it was about dosing and drug-supply contamination, exactly the kind of granular, practical concern you’d expect from people who’ve actually used the tool under pressure, not from people debating it in the abstract. That’s a useful reminder for anyone designing the next round of harm reduction policy: the population closest to the intervention isn’t asking whether it works. They’re asking how to make it work better.

That’s a useful reminder for anyone designing the next round of harm reduction policy: the population closest to the intervention isn’t asking whether it works.