Overdose Deaths Are Falling. Most People Who Need Help Still Never Reach Treatment.

Something unusual happened in American public health over the past three years, and almost nobody threw a party.

Drug overdose deaths fell. Then fell again. Then fell a third time. Provisional figures from the Centers for Disease Control and Prevention put the 2025 total at roughly 69,973 — down close to 14 per cent from the 81,313 estimated for 2024, and back to a level last recorded before the pandemic. Deaths involving opioids dropped from an estimated 55,296 to 44,564. Almost every state improved, several by more than a quarter.

It is the longest sustained decline in decades. On paper, it looks like a turning point.

Inside treatment centres, it mostly doesn’t feel like one.

The reason sits in a second set of numbers that has barely moved. In the federal government’s National Survey on Drug Use and Health, only about one in five people who needed substance use treatment in 2024 actually received any. Four in five did not. That ratio has been stubborn for the better part of two decades, through funding surges, policy shifts and three separate national conversations about the opioid crisis.

Which produces an uncomfortable reading of the good news. Fewer people are dying. Roughly the same proportion are getting well.

Staying alive is not the same as getting better

Much of the credit for the decline goes to naloxone — the overdose reversal medication now carried by paramedics, police officers, librarians and a growing number of ordinary people who have learned to expect the worst. It works. It is one of the most cost-effective interventions in modern medicine.

It also does nothing about Tuesday.

Naloxone interrupts a death. It does not touch the reason someone was using alone in the first place, and it does not book anyone an appointment. Shifts in the illicit drug supply have helped too, along with wider access to medications for opioid use disorder. All of it keeps people breathing. None of it, by itself, changes the trajectory of an illness that tends to reassert itself the moment the emergency passes.

Clinicians describe this as the gap between rescue and repair. The rescue system in the United States has improved enormously. The repair system has not kept pace.

The window nobody plans for

Anyone who has watched a family member cycle through crisis will recognise the pattern.

The emergency happens. Everyone agrees, out loud and with real feeling, that something has to change. Calls are made. And then, somewhere in the two or three days it takes to verify insurance, find an available bed and arrange time away from work, the urgency quietly drains out of the room. The person who was ready on Sunday has usually stopped being ready by Wednesday.

That window — the short, unreliable stretch between willingness and admission — is where most of the treatment gap actually lives. It rarely appears in policy documents, because it is not a funding problem or a capacity problem in any clean sense. It is a timing problem. Motivation in addiction is not a stable resource. It arrives in bursts, usually after something frightening, and it fades faster than most systems can respond.

Providers who understand this build for speed rather than elegance: same-day assessment, insurance verified while the person is still on the phone, transport arranged before anyone has time to reconsider. It is unglamorous work. It is also, arguably, the single highest-leverage improvement available.

Most people do not think they need help

There is a second reason the ratio stays flat, and it is less about systems than about self-perception.

The overwhelming majority of people who meet the clinical criteria for a substance use disorder do not believe they need treatment. This is not denial in the theatrical sense — the shouting, the slammed door, the intervention scene from a film. It is quieter and more reasonable-sounding than that. People compare themselves to someone worse. They point to the job they still have, the mortgage they still pay, the mornings they still manage. They are not lying. They are measuring themselves against a caricature of addiction that almost nobody actually matches.

Federal survey data suggests the underlying population is shifting slightly in the right direction. Past-year substance use disorder among people aged 12 and over fell from 16.7 per cent in 2021 to 15.3 per cent in 2025, with alcohol use disorder dropping from 10.6 per cent to 8.9 per cent over the same period. Adolescent indicators improved across several measures.

But a smaller pool of people who need care does not help much if the proportion reaching it stays where it is.

The part the numbers hide

There is a structural problem underneath all of this that the overdose statistics cannot show.

Substance use and psychiatric illness travel together far more often than the treatment system is organised to handle. Depression, anxiety, trauma histories, bipolar disorder — these are not side notes in addiction cases. They are frequently the thing that started it. Yet mental health services and addiction services in most of the country are separately funded, separately staffed and separately regulated, which means a person with two conditions is routinely asked to solve them one at a time, in sequence, in different buildings.

It seldom works. Treating the drinking without treating the panic disorder tends to produce a relapse that everyone involved then interprets as a failure of willpower.

Programs built around dual diagnosis take the opposite approach, running psychiatric assessment and medication management alongside addiction care from the first week rather than the fourth month. Providers such as First Step Behavioral Health, a Joint Commission-accredited dual-diagnosis facility in Pompano Beach, Florida, structure care as a continuum — detox, residential, partial hospitalisation, intensive outpatient, aftercare — on the reasoning that discharge is where most recoveries are actually lost.

That reasoning is well supported. The riskiest period for someone leaving residential care is the fortnight immediately after, when structure disappears overnight and the old environment is waiting exactly where it was left.

What families should actually ask

For anyone comparing one addiction treatment facility against another, the useful questions are narrower than the marketing suggests, and none of them concern the swimming pool.

How quickly can someone be admitted, in days? Is there a psychiatrist on site, and how often is a patient seen? Are medications for opioid or alcohol use disorder available, or does the programme discourage them? What happens in the fortnight after discharge — is there a named plan, or a leaflet? Is the facility accredited, and by whom? And what proportion of patients complete the programme, rather than what proportion say they enjoyed it?

Facilities that answer those questions directly are rarely the ones with the largest advertising budgets.

The number worth watching

The overdose figures will probably keep improving for a while. Naloxone distribution is still expanding, and the drug supply appears to be shifting.

That progress is real and it is worth defending. But it measures how well the country has learned to interrupt a death, which is a different achievement from helping someone build a life that no longer requires interrupting.

Four in five. Until that figure moves, the crisis has not turned. It has only slowed down enough to be ignored.

---

Anyone struggling with substance use or mental health can call or text 988 in the United States, or contact SAMHSA’s national helpline on 1-800-662-HELP (4357), which is free and available around the clock.