Something genuinely good has been happening, and families deserve to hear it stated plainly before anything else.
Drug overdose deaths in the United States have fallen for three consecutive years. Provisional data from the CDC’s National Center for Health Statistics estimates roughly 69,973 overdose deaths in 2025, down almost fourteen percent from an estimated 81,313 in 2024. Deaths involving opioids fell from an estimated 55,296 to 44,564 over the same period. Almost every state saw a decrease.
The year before that was larger still. A peer-reviewed analysis of CDC mortality records, published in the journal Addiction in June 2026 by researchers at the University of California San Diego, found the overdose death rate fell 24.4 percent between 2023 and 2024, the first recorded decline across all four waves of the overdose crisis.
That is tens of thousands of people alive who would not have been on the previous trajectory. It is not a statistical artifact and it should not be minimized.
But we have started getting a particular question from families in Houston, and it deserves a careful answer rather than a reassuring one. The question is some version of: if things are getting better, do we have more time?
The answer is no, and the reasons are worth understanding.
What the decline covers, and what it does not
National overdose totals are driven overwhelmingly by opioids, primarily illicitly manufactured fentanyl. When fentanyl deaths fall sharply, the total falls sharply, because that category is most of the total.
The Addiction analysis shows exactly that. Fentanyl-involved deaths without stimulants dropped from 31,193 in 2023 to 19,673 in 2024. Fentanyl-involved deaths that also involved stimulants dropped from 41,583 to 28,062. Those two categories account for nearly the whole improvement.
The CDC also reported that cocaine and psychostimulant deaths decreased in 2025. That is worth stating clearly, because a great deal of writing on this subject assumes the opposite.
There is a wrinkle underneath it, though, and it is the part that matters for families.
A large share of stimulant-involved deaths also involve fentanyl. When a death involves both methamphetamine and fentanyl, it is counted in both categories. So when the opioid supply becomes less lethal or less prevalent, deaths fall in the stimulant column too, without anything necessarily having improved about stimulant use itself.
Separate out the deaths involving stimulants and no fentanyl, and the direction reverses. The same analysis found that category rose from 18,142 deaths in 2023 to 18,907 in 2024, while everything else was falling. And because the overall total dropped so far, those deaths grew from 17.3 percent of all overdose deaths to 23.8 percent. Close to a quarter.
So the honest summary is this. The crisis is easing, substantially, and it is easing unevenly. The part improving fastest is the part driven by fentanyl. The part not improving is the part driven by stimulants on their own, and it is now a much larger share of what remains.
Why that matters if methamphetamine or cocaine is the issue in your family
Three practical consequences follow, and none is obvious from a headline.
There is no approved medication for stimulant use disorder. For opioid use disorder, approved medications exist with substantial evidence behind them, and the national response has been built around those medications alongside naloxone distribution. Nothing equivalent exists for methamphetamine or cocaine. The National Institute on Drug Abuse states plainly that no FDA-approved medication exists for methamphetamine use disorder or any other stimulant use disorder, and the FDA said the same in announcing draft guidance intended to help developers bring such treatments forward. That gap is a substantial part of why the stimulant trend has not followed the opioid one.
That does not mean treatment is unavailable, and this is the part families most often have backwards. NIDA describes effective behavioural treatments for stimulant use disorder, and identifies contingency management as the best studied and the one most associated with treatment success. It works by providing small, tangible incentives for continued abstinence and continued engagement in treatment. Cognitive behavioural therapy, motivational interviewing, and group support are also named as evidence-based approaches.
This is worth knowing before you speak to any program. If a stimulant is the presenting problem, a reasonable question to ask is what the program’s approach actually consists of, and whether contingency management is part of it. Families who do not know to ask often assume the absence of a medication means the absence of a plan, and it does not.
Naloxone does not reverse a stimulant overdose. It reverses opioid overdose. Because fentanyl so often turns up in stimulant supplies, naloxone should still be on hand and should still be used if someone is unresponsive, since there is no way to know at the time what is involved. But a family who has naloxone and believes they are therefore prepared may not be as prepared as they think.
There is a related gap worth closing this week. Someone who has never sought an opioid, and does not consider themselves an opioid user, frequently has no naloxone, has never discussed overdose with anyone, and has no plan. We have written separately about cocaine and methamphetamine overdose risk in Houston, and this is the widest hole in it. Closing it requires no conversation about addiction and no decision about treatment.
Why a falling national rate says nothing about one person
This is the core of it.
A national statistic describes a population. It tells you how many households got the worst phone call last year compared with the year before. It contains no information about whether a particular person is safer than they were.
Two things in particular make individual risk poorly correlated with the national trend.
The supply is variable, not uniformly weaker. A falling death rate is entirely consistent with an unpredictable supply in which some batches are far stronger than others. A person whose tolerance is calibrated to what they have been getting, who encounters something considerably stronger, is in acute danger. That variance does not appear in an annual death count, which averages it away.
Geography diverges from the national picture. The CDC’s provisional 2025 data showed almost all states declining, but the spread was wide. Some fell by more than a quarter. New Mexico, Arizona and Colorado rose by ten percent or more. A national figure describes the country, not any particular place inside it, and the same is true of any single county within a state.
What this means for the timing of a decision
Families ask whether a falling rate buys them more time. It is a reasonable question, and the answer is that it buys nothing at the household level.
The reasoning that leads there is the same reasoning behind waiting for rock bottom, arriving in a more sophisticated form. Both treat a future event as something the family can plan around: a crisis that will provide clarity, or a trend that will provide safety. Neither is under anyone’s control, and neither is a plan.
Rock bottom in particular is not a clinical concept. The evidence on family-directed approaches has moved firmly away from it. Reviews of the Community Reinforcement and Family Training model, developed for families of someone refusing help, explicitly identify “wait until they hit rock bottom” and “detach and take care of yourself” as the guidance families were given historically in the absence of anything better. A meta-analysis of studies of that model found it roughly twice as effective at engaging a person in treatment as the comparison approaches tested against it, with engagement rates varying considerably across individual studies.
The useful conclusion is not that the good news is fake. It is that the good news is about the country, and your decision is about a person.
How Crosswell Interventions helps
Crosswell Interventions provides clinician-led intervention services to families across the United States, including as a drug and alcohol interventionist in Houston and throughout Texas. Will Crosswell is a Licensed Clinical Social Worker and Licensed Chemical Dependency Counselor and a Certified Clinical Interventionist trained by Debra and Jeff Jay at Love First.
Where polysubstance use is involved, the Crosswell Method’s five steps carry specific weight.
Listening establishes what is actually being used, all of it, which families frequently do not know in full. Different members of the household have different pieces of this.
Invitation is why this is not a confrontation. The goal is a person accepting help rather than being cornered into agreeing.
Planning is where polysubstance use changes everything. What is being used determines whether a medical assessment is needed first, what level of care is appropriate, and whether a given program can accept the person at all. A yes with no suitable destination confirmed tends to become a no within a day.
Treatment transition covers the gap between agreement and arrival, including sober transport.
Ongoing family support continues through regular family sessions and clinical case management.
We coordinate treatment placement, and through our partnership with Kind we can integrate psychiatric care where mental health needs sit alongside the substance use. We are reachable around the clock.
Frequently asked questions
Are overdose deaths going down in Texas? Nationally, yes. CDC provisional data shows a third consecutive annual decline, with almost every state decreasing. For current Texas and Harris County figures, the Texas Department of State Health Services maintains a public dashboard, and county-level medical examiner data is the most local source available. National direction and local reality can differ, so it is worth checking the local number rather than assuming.
Does naloxone work if the overdose involves methamphetamine? Naloxone reverses opioid overdose. It does not reverse the effects of methamphetamine or cocaine. It should still be kept and still be used if someone is unresponsive, because fentanyl frequently turns up in stimulant supplies and there is no way to know what is involved at the time. Call 911 either way.
If there is no medication for methamphetamine or cocaine addiction, what does treatment involve? Behavioural treatment. The National Institute on Drug Abuse identifies contingency management, which provides small tangible incentives for continued abstinence and engagement, as the best studied approach and the one most associated with treatment success. Cognitive behavioural therapy, motivational interviewing and group support are also evidence-based. When a stimulant is the presenting problem, it is worth asking a program directly which of these they use.
If the numbers are improving, can we wait a little longer to act? The improvement is real and it describes a population, not a person. An unpredictable supply means individual risk can be high in a year when national deaths fall. A falling rate does not extend anyone’s runway, and waiting for a trend to make a decision safer works about as well as waiting for a crisis to make it obvious.
The thing worth remembering
Fewer families are getting the worst phone call than three years ago. That is true, it is significant, and nobody should talk anyone out of taking some comfort in it.
Whether yours is one of them was never a question the national numbers could answer.
