Notes from the work.

A sober companion is a trained professional who stays physically present with someone in early recovery through a defined high-risk window: a flight home from treatment, a work conference, a wedding, the first weeks back in the house. Not a bodyguard, not a nurse, and not supervision. Company, with a plan. The job is to make sure the hardest hours of early recovery are never faced alone.

What the work actually looks like

The days are deliberately unremarkable. A companion on a business trip gets the client to the airport lounge without a stop at the bar, sits through the dinner where everyone else is drinking, and is in the next seat, or the next room, when the 11 p.m. restlessness hits. During a reentry week at home, it’s morning structure, meetings attended together, and a calm presence when the first argument with a spouse lands harder than expected.

The skill isn’t dramatic rescue. It’s noticing the drift two hours early: the changed plan, the sudden errand, the phone that goes face down. People rarely lose their footing in treatment. They lose it in the gaps between, and a companion’s whole job is standing in a specific gap for as long as it stays open.

Timing explains why this works. Early recovery is a stretch where cravings still fire on cue while the brain’s braking system is still coming back online; we walk through that science in what addiction does to the brain. A companion is borrowed impulse control for the exact hours it’s most likely to be overmatched.

Companion, transport, coach: which one is this?

Families hear these terms used loosely, including by providers. The real distinction is scope.

Sober transport is a journey. Door to door, one direction: home to detox, residential to airport to home. It ends at the destination. Companion support is a window. Days to weeks of continuous or near-continuous presence through a known risk period. Recovery coaching is a relationship. Weekly structure, relapse-risk planning, and escalation support over months, mostly by phone and in scheduled sessions rather than constant presence; our page on what a recovery coach does covers it properly.

In a well-built plan they hand off to each other: transport gets the person home from treatment, a companion covers the first two weeks of reentry, and a coach carries the next year. The evidence on continuing care favors exactly this shape, support that lasts a year or more and actively stays engaged rather than waiting for a call (McKay, 2021). One firm covering all three also means one person who knows the whole story. No handoffs to strangers, no re-explaining.

When it’s worth it, and when it isn’t

The honest version: a companion is a targeted tool, not a lifestyle. The cases where it earns its cost share one feature, a known window of elevated risk with a start and an end. Travel that can’t be postponed, an event dense with old cues, the first stretch home after residential when the house is full of Saturdays and nothing is scheduled. Our note on business travel in early recovery walks through the most common one.

What a companion can’t do is substitute for treatment or for the longer structure around it. Presence without a plan is just an expensive roommate.

What to ask before you hire one

Four questions sort the field quickly. What’s your training, and are you in recovery yourself? What exactly is the scope: hours, nights, travel, and what happens if there’s a crisis at 2 a.m.? How do you communicate with the family, and with the clinical team? And who do you escalate to, by name, if the plan stops working? Vague answers to the last two are the tell. A professional companion is part of a team, and says so. If you’re weighing whether a companion, a coach, or a driver is what your situation actually calls for, that’s a conversation we have with families every week. Call and ask; a person answers.

An intervention is a planned, structured conversation in which the people closest to someone ask them to accept treatment, with the treatment already arranged. That last clause is most of the definition. The television version is an ambush with speeches. The professional version is two weeks of preparation for a conversation that takes forty minutes, with a bed reserved before anyone sits down.

What actually happens

A prepared intervention runs in phases. First, planning with a professional: who should be in the room, and just as important, who shouldn’t. Letters are drafted and rehearsed; specific, loving, free of blame. Consequences are agreed in advance, which means each family member decides what they will actually do if the answer is no, and commits to it out loud. Treatment is arranged completely: program chosen, admission confirmed, transport planned, bag packed. Then the conversation itself, which is short, calm by design, and aimed at a single question with a yes or no answer.

The preparation is the intervention. The meeting is just where it becomes visible.

Does it ever fail? Honestly, yes

Two different things get called failure, and they deserve different answers.

The first is the family that never gets to the room. In the randomized trial that tested the confrontational intervention against other family approaches, the striking finding wasn’t the refusal rate of the person using. It was that most families assigned to the confrontational method chose not to go through with the meeting when the moment came (Miller, Meyers and Tonigan, 1999). Asking a family to stage a surprise confrontation with someone they love turns out to be asking a great deal. Overall, about 30 percent of that group’s loved ones entered treatment, against 64 percent for CRAFT, a lower-drama family training method we’ve written about in our post on boundaries and family programs. If nobody has told you there’s an evidence-backed alternative to the ambush, now someone has.

The second kind of failure is a no in the room. It happens, and a well-run intervention prices it in. The agreed consequences begin, the door stays open, and in a meaningful share of cases the yes arrives days or weeks later, when the new reality settles. A no is a delay, not a verdict, provided the family holds the line they rehearsed.

So the honest summary: interventions work often enough to be worth doing properly, they are not the highest-yield first move for every family, and the failure mode that is actually avoidable is poor preparation. Which brings up the part almost nobody plans for.

The hour after yes

Yes is not the finish line; it’s the most fragile hour of the whole process. The person who agreed at 10 a.m. is negotiating by noon: one more night at home, a stop to say goodbye to someone, I’ll fly out Monday. Every added hour between yes and admission is an hour for fear to renegotiate.

This is why the logistics belong inside the plan, not after it. Bed confirmed before the meeting. Bag already in the car. And someone whose job is the journey itself: our sober transport service exists precisely for this window, a trained companion door to door, across town or across the country, so the distance between yes and intake is measured in hours and none of them are unaccompanied.

For a fuller walk-through of the sequence, our guide to what an intervention looks like covers the planning in detail, and our intervention support page explains how we work with families before, during, and after the room. Whatever route you take, take the prepared one. The conversation is too expensive to improvise.

Sooner or later a doctor says the word Suboxone, and the family goes quiet. Then somebody says the sentence: isn’t that just trading one addiction for another?

We are not clinicians. We do not prescribe, and nothing here is medical advice; that conversation belongs to the person and their doctor. But families are routinely asked to form an opinion on this without ever being shown the evidence, and the evidence is not close. So here it is, in plain terms.

What the medications are

Three are used for opioid addiction.

Methadone. A long acting opioid, taken once a day, usually at a licensed clinic. It settles onto the same receptors the drug used, but steadily, with no rush and no crash. Withdrawal stops. Craving drops a lot. The person can hold a job and sleep at night.

Buprenorphine (brand names Suboxone, Sublocade). Also an opioid, but a partial one. It has a ceiling: past a certain dose, more does not produce more effect, which is why it is much harder to overdose on. It can be prescribed in a regular doctor’s office, and there is a monthly injection version.

Naltrexone (brand name Vivitrol, also a daily pill). Not an opioid at all. It blocks the receptors, so opioids stop working. It has one catch, and it is a big one: a person has to be completely through withdrawal before starting, or the blocker throws them into it.

Two are used for alcohol: naltrexone again, and acamprosate.

The data is not really about comfort. It is about dying.

The largest analysis of this question pooled cohort studies following people in and out of treatment. Across 16 cohorts and 122,885 people on methadone, the all cause death rate was 11.3 per 1,000 person years while people were in treatment and 36.1 while they were out of it. Overdose deaths were 2.6 in treatment against 12.7 out. For buprenorphine, across 3 cohorts and 15,831 people, all cause deaths were 4.3 in treatment against 9.5 out (Sordo et al., The BMJ, 2017).

Read those pairs again. Being on the medication was associated with roughly a third of the death rate of being off it.

The same analysis found the two most dangerous stretches are the first four weeks after starting and the first four weeks after stopping. Stopping is the risky part, not staying.

Then there is the study families should probably see first. Researchers followed 40,885 people with opioid use disorder through six different treatment paths and compared what happened next. Only one path was associated with fewer overdoses: buprenorphine or methadone, with about a quarter of the overdose risk of no treatment at three months and about 40 percent of it at twelve (Wakeman et al., JAMA Network Open, 2020). Inpatient detox, residential treatment, intensive outpatient, counseling, and naltrexone did not show a significant reduction in overdose in that data.

That finding is uncomfortable, and we are not going to soften it. The 30 day residential admission is the thing most families are sold, sometimes at very high cost, and in this study it was not the thing that moved the overdose number. It does other work. It is not this.

One honest caveat: these are observational studies, not coin flips. People who stay on medication may differ from people who do not in ways the researchers could not fully adjust for. The findings are large, consistent, and repeated across countries, which is why they carry weight, but they are an association, not a guarantee.

About the substitution objection

Here is the fair version of it and the honest answer.

A person on a steady dose of methadone or buprenorphine is not high. The dose is level, so there is no peak to chase and no crash to survive. They can drive, work, parent, and remember the conversation. What changed is the shape of the exposure and, in the data above, whether they live.

But the objection is not silly. These are opioids. Stopping them causes withdrawal. Some people want to be off everything eventually, and that is a legitimate goal, not a failure of commitment. The research does not say never stop. It says the exit is the dangerous part, so it should be slow, planned, and supported, not improvised in a bad week.

If naltrexone is the goal, know the hurdle

Families often prefer naltrexone, because a blocker feels cleaner than an opioid. A large trial compared the monthly naltrexone shot against buprenorphine in 570 people. Twenty eight percent of the naltrexone group never managed to start the medication at all, against 6 percent of the buprenorphine group, because getting fully through withdrawal first is genuinely hard. Counting everyone assigned, buprenorphine came out ahead. But among the people who did get started on each one, the results were essentially the same (Lee et al., The Lancet, 2018).

So naltrexone works about as well, for the people who can get on it. Which means the practical question is where the person will be during that withdrawal window, and who is with them.

Alcohol has medications too, and almost nobody is offered them

For alcohol, a review of dozens of trials found that for every 12 people treated with acamprosate, one additional person avoided returning to drinking. For oral naltrexone at 50 mg a day, for every 12 treated, one additional person avoided returning to heavy drinking (Jonas et al., JAMA, 2014).

That is a modest effect and a real one. It is also rarely mentioned. If a family member is in treatment for alcohol and no one has raised medication, that is worth asking about.

The number that should bother everyone

Researchers tracked 17,568 adults in Massachusetts who survived an opioid overdose. In the twelve months afterward, 11 percent received methadone, 17 percent received buprenorphine, and 6 percent received naltrexone. For those who got methadone or buprenorphine, death rates were roughly cut in half (Larochelle et al., Annals of Internal Medicine, 2018).

Turn that around. Most people survive an overdose and are sent home with nothing.

Three questions to ask any program

You do not need to pick a medication. That is the doctor’s job and the person’s choice. But you can ask these, and the answers tell you a lot.

Does this program allow medication, or does it require people to be off it? Some abstinence based programs and many sober homes will not accept a resident who is taking buprenorphine. Ask before the deposit, not after.

Who prescribes it after discharge, and is that appointment already booked? A prescription that lapses in the gap between levels of care is the same as no prescription, and that gap is where plans break. We wrote about that seam in the gap between levels of care.

If the person ever chooses to come off it, what is the plan for the month after? The data says that month is the most dangerous one in the whole sequence. Nobody should reach it without a plan.

Where we fit

We are not a medical provider and we do not manage medication. What we do is keep the plan from falling apart between the people who do.

In practice that means recovery coaching supplying the weekly rhythm and making sure the appointment actually gets kept, and family coordination meaning one person is tracking whether the prescription got refilled while the treatment team changes around it. If you are still choosing a program, our guide to choosing a treatment center has the medication question built into the tour list. And if the brain science behind all of this is what you want next, what addiction does to the brain covers why the first year needs borrowed structure.

None of this settles what your family should do. It just means the decision gets made with the numbers in the room, which is more than most families are given.

Sources

• Sordo L, Barrio G, Bravo MJ, et al. Mortality risk during and after opioid substitution treatment: systematic review and meta-analysis of cohort studies. The BMJ, 2017;357:j1550. bmj.com/content/357/bmj.j1550
• Wakeman SE, Larochelle MR, Ameli O, et al. Comparative effectiveness of different treatment pathways for opioid use disorder. JAMA Network Open, 2020;3(2):e1920622. jamanetwork.com
• Lee JD, Nunes EV, Novo P, et al. Comparative effectiveness of extended-release naltrexone versus buprenorphine-naloxone for opioid relapse prevention (X:BOT). The Lancet, 2018;391(10118):309-318. pmc.ncbi.nlm.nih.gov/articles/PMC5806119
• Jonas DE, Amick HR, Feltner C, et al. Pharmacotherapy for adults with alcohol use disorders in outpatient settings: a systematic review and meta-analysis. JAMA, 2014;311(18):1889-1900. bumc.bu.edu
• Larochelle MR, Bernson D, Land T, et al. Medication for opioid use disorder after nonfatal opioid overdose and association with mortality: a cohort study. Annals of Internal Medicine, 2018;169(3):137-145. acpjournals.org/doi/10.7326/M17-3107

Families meet the treatment system mid-crisis, which is the worst possible moment to learn a new vocabulary. Detox, residential, PHP, IOP: admissions staff use these terms as if everyone was issued a glossary. Here is the glossary, in plain language, plus the one thing about the system nobody explains on the phone.

The framework behind the jargon comes from the American Society of Addiction Medicine, whose criteria match a person’s medical and life situation to an intensity of care, stepping down as stability builds (the ASAM criteria).

The levels, top to bottom

Medical detox (withdrawal management). Days, not weeks. Its only job is getting someone through withdrawal safely, with medical staff and often medication. Alcohol and benzodiazepine withdrawal can be dangerous and belong under medical supervision. What detox is not: treatment. A person leaving detox has a stabilized body and an untouched addiction.

Residential treatment. Usually 28 to 90 days living at the facility. Structure all day: individual and group therapy, a fixed schedule, no access. This is what most people picture as “rehab.” Its strength is containment; its weakness is that containment ends.

PHP, partial hospitalization. Treatment five or six days a week, most of the day, but the person sleeps elsewhere: at home, or in structured housing. Real therapy hours, real-world evenings.

IOP, intensive outpatient. Roughly nine to fifteen hours a week, often in evening blocks, built to run alongside work or school. The person’s life is mostly back in their own hands.

Outpatient and aftercare. An hour or two a week of therapy, plus whatever recovery community the person builds. From here on, the schedule is theirs.

Each step down transfers responsibility from the program to the person, which is the point. Nobody recovers into a locked schedule; they recover into a Tuesday with three meetings and an empty evening.

The seams are the risk, not the levels

Now the part the brochure leaves out. Every one of those transitions is a handoff, and the handoffs are where plans break. Fewer than a third of adults leaving inpatient or residential care have any treatment contact within the next seven days; we walked through that data in the gap between levels of care. The levels are staffed. The seams between them belong to nobody, unless somebody is given them by name.

In practice the seams look small. A discharge on Friday with an IOP intake the following Thursday. A flight home from a Utah program, alone, with a six-hour layover. A car ride from residential to PHP housing that passes the old neighborhood. Each one is a few unsupervised hours at the exact moment tolerance is down and the brain’s cue response is still loud.

This is the specific problem our sober transport service exists to remove: a trained companion, door to door, so the highest-risk hours of the whole sequence are never unaccompanied. For the trip out of residential in particular, we’ve written about why the ride home deserves more planning than it gets. And across the longer seams, the weeks between IOP and ordinary life, a recovery coach keeps one person accountable for continuity while the clinical team changes at every level.

Three questions to ask any program

You don’t need to master the criteria. Ask these instead. What level is this, and what level comes next? Is the next level booked before discharge, with a date? And who, by name, is responsible for my person between the two? A good program answers all three without flinching. If the third answer is “the family,” you now know which job is yours to fill, or to hand to someone who does it professionally.

If you’ve watched someone you love choose a substance over everything they used to care about, the question underneath everything is usually some version of: why can’t they just stop? The brain science has a real answer. It won’t make the choices hurt less, but it explains them, and it tells you what kind of help the first year actually requires.

The circuit that decides what matters

Deep in the brain there’s a pathway, called the mesolimbic dopamine system, whose job is to tag things as worth pursuing: food, connection, accomplishment. It runs on dopamine surges, and it learns. Substances of abuse hit this circuit far harder than any natural reward, and with repetition the brain adapts in two directions at once. It becomes exquisitely sensitive to cues that predict use: the exit off Federal Highway, a payday, a particular friend’s name on the phone. And it turns the volume down on everything else, so ordinary rewards land flat (Volkow, Koob and McLellan, New England Journal of Medicine, 2016).

That second part matters more than families expect. Early recovery isn’t neutral. It’s a stretch where the brain’s reward system registers less pleasure from everything, while cues still fire hard. The clinical term is anhedonia. The lived version is a person, three weeks out of treatment, who says everything feels gray and can’t explain why.

The circuit that hits the brakes

The second change happens up front, in the prefrontal cortex, the region that weighs consequences and vetoes bad ideas. Imaging studies reviewed in the same NEJM paper show reduced activity there in people with long-term addiction. The craving circuit gets stronger; the braking circuit gets weaker; and the collision of those two facts is what “loss of control” means physiologically. It isn’t an empty will. It’s an overmatched one.

This is why lectures and consequences alone change so little in the first months. The argument is being addressed to the part of the brain that’s currently offline.

The part families rarely get told: it comes back

The changes are not permanent. A study using brain imaging in methamphetamine users found that dopamine transporter levels, sharply reduced at the start of abstinence, showed substantial recovery after roughly 14 months (Volkow et al., Journal of Neuroscience, 2001). Prefrontal function improves along a similar arc. The brain recovers on the timescale of a year, not a weekend.

Read that timeline against what most plans provide: 30 days of structure, then a calendar full of nothing.

Borrowed structure, on purpose

The practical conclusion is almost mechanical. For something like a year, the judgment and impulse control the person will eventually supply themselves has to be borrowed from outside. Not as punishment, as scaffolding while the wiring recovers. That’s the design logic behind what we do: a recovery coach supplies the weekly rhythm and the escalation plan while the prefrontal cortex gets back online, and companion support covers the specific hours when cues are loudest: the business trip, the wedding, the first weeks home. Our resource on the first ninety days maps what that structure looks like week by week.

None of this is destiny. It’s a repair schedule. The brain that learned addiction can unlearn it, with time, protected from the windows where the old circuit still outvotes the new one. The plan just has to be built for the brain the person has this year, not the one they’ll have next year.

The short answer: yes. The American Medical Association classed alcoholism as an illness in 1956, addiction followed, and every major medical body since has held the line. But families don’t ask the question because they want a vote count. They ask because the answer decides what a reasonable plan looks like, and that’s where the disease model earns its keep.

The comparison that reframes everything

In 2000, researchers led by A. Thomas McLellan published a comparison in JAMA that is still the clearest way into this. They set drug dependence beside three conditions nobody argues about: type 1 diabetes, hypertension, and asthma. Genetic contribution, similar. The role of behavior and environment in how the condition runs, similar. Adherence to treatment, similar, and similarly imperfect (McLellan, Lewis, O’Brien and Kleber, JAMA, 2000).

Then the numbers that matter to a family staring at a relapse. Roughly 40 to 60 percent of people treated for a substance use disorder return to use within a year. For hypertension and asthma, 50 to 70 percent of adults have a recurrence of symptoms each year that requires additional care. Diabetes sits in a similar band.

Nobody concludes from an asthma attack that the inhaler was a scam or that the patient didn’t want it enough. The same event in addiction gets read as moral failure, by families and often by the person themselves. The data doesn’t support that reading.

What “chronic” actually asks of you

Here is the practical weight of the word. Chronic conditions aren’t cured in an episode of care; they’re managed over years, and management is mostly unglamorous structure. A diabetic doesn’t leave the endocrinologist cured. They leave with a monitoring routine, a person to call, and a plan for the bad week.

Addiction handled the same way looks like this: treatment as the acute phase, then a long stretch of maintenance that somebody has to own. Regular contact with a person who knows the whole story. A written plan for the high-risk windows: the business trip, the wedding, the first month back at work. Escalation before a slip becomes a run. That maintenance layer is what recovery coaching is, and the research on continuing care backs the shape of it: the programs that work are the ones that last a year or more and actively pursue the person, rather than waiting by the phone (McKay, 2021).

What the disease model is not

It is not a free pass. A diagnosis explains behavior; it doesn’t excuse the work of managing it. The person still has to do the daily part, and the family still gets to hold boundaries. Those two things live comfortably inside the disease frame, and we’ve written about the strongest evidence-based way for families to do their half in our post on family programs and boundaries.

It’s also not a prognosis. “Chronic” describes how the condition behaves untreated, not how the story ends. The same long-term research that produces the scary relapse figures also shows that outcomes climb steadily with sustained engagement; we walk through those numbers in what the long-term data actually says.

So when someone asks us whether addiction is really a disease, we say yes, and then we ask the more useful question: if you treated this exactly as seriously as diabetes, what would next month look like? Book the follow-up before the discharge. Name the person who owns the quiet weeks. That’s the whole answer, applied.

If someone you love is using and refusing help, you are not a bystander in this. That is not encouragement. It is a finding. What the family does is one of the most reliably measured variables in the whole field, and the approach with the strongest evidence behind it is the one almost nobody tells families about.

The trial that should be better known

In 1999, Miller, Meyers and Tonigan published a randomized comparison in the Journal of Consulting and Clinical Psychology of three ways a family member could try to get an unmotivated problem drinker into treatment. 130 concerned significant others were randomly assigned.

The results, by treatment entry:

• Community Reinforcement and Family Training (CRAFT): 64 percent
• Johnson Institute intervention, the confrontational family meeting: 30 percent
• Al-Anon or Nar-Anon facilitation: 13 percent

(Miller, Meyers and Tonigan, 1999)

A later review in Addiction pooled four randomized trials covering 264 family members and found CRAFT produced a relative risk of engagement of 3.25 (95 percent CI 2.11 to 5.02) against twelve-step facilitation, with a number needed to treat of two, and 2.15 (95 percent CI 1.28 to 3.62) against the Johnson intervention, number needed to treat of three (Roozen, de Waart and van der Kroft, 2010). A number needed to treat of two means roughly one additional person enters treatment for every two families trained.

The trials are small and the authors say so. But the direction has held across every replication, and it points somewhere counterintuitive.

What CRAFT is, and what it is not

CRAFT trains the family, not the person using. It teaches how to reinforce non-using behavior, how to withdraw from conflict rather than escalate, how to stay safe, how to recognize and use the windows when someone is briefly open to help, and how to ask in a way that can be said yes to. It is not confrontation, and it is not detachment. It is a third option that most families are never offered because the two they have heard of occupy all the airtime.

Notice which approach came second. The surprise intervention, the version everybody has seen on television, engaged 30 percent. Real, and less than half of CRAFT. If you have been told that confrontation is the only remaining move, the research does not support that. Interventions have a place, particularly where safety is immediate, and they work better when they are planned with the family rather than performed on them. They are not the highest-yield first step.

In defence of Al-Anon, honestly

Thirteen percent is a poor engagement rate, and it is also the wrong measure. Al-Anon does not claim to get anyone into treatment. Its stated purpose is the wellbeing of the family member, and judging it on a goal it never set is unfair.

On its own terms it holds up. Across those trials, all three approaches produced improvement in the family member’s own functioning at six months, measured on depression, anger, family cohesion, relationship happiness and family conflict (CCSA summary of the CRAFT evidence). The CRAFT trials showed significant improvement in depression symptoms on the Beck Depression Inventory over six months.

So the sensible reading is not that one replaces the other. It is that they answer different questions. If your question is how to get someone into treatment, the evidence points to CRAFT. If your question is how to survive the next two years without losing yourself, Al-Anon has decades of people who will tell you it did that job. Most families need both answers.

Boundaries, defined usefully

Families are told to set boundaries constantly and told what one is almost never. A working definition:

A boundary is a statement about your own behavior, not a demand about theirs. “You have to stop using” is not a boundary; it is a wish, and it hands the outcome to the person least able to deliver it right now. “We will not give cash, and we will not pay for the car if there is a positive test” is a boundary, because you control both halves.

Four rules that make them hold.

• Enforceable by you alone. If keeping it requires their cooperation, it is not a boundary.
• Written down. Verbal boundaries drift under pressure, and pressure is the condition they exist for.
• Agreed between the adults first. The most common failure is not the boundary. It is one parent quietly funding around the other. Split households get worked, every time, and nobody has to be malicious for it to happen.
• Few. Three or four you will actually hold beats fifteen you will not. Every boundary you set and abandon teaches that the next one is negotiable.

And the part families flinch at: a boundary is not a punishment and it is not leverage. It is the description of a relationship you can sustain without being destroyed by it. Framing it as a threat invites a negotiation. Framing it as a fact does not.

What we do here

Alongside is a non-clinical service, and we do not diagnose or treat. We work alongside the family: helping decide what the boundaries are before the conversation rather than during it, deciding who says what, keeping the adults aligned when the pressure arrives, and staying present through the weeks afterward, whichever way it goes.

Planned in days, not weeks; with the family, not around them.

If you are somewhere in this, speak with an advisor. For the conversations themselves, Scripts for families gives you actual language, and What an intervention looks like sets out how one is planned when it is the right call.

Sources

• Miller WR, Meyers RJ, Tonigan JS. Engaging the unmotivated in treatment for alcohol problems: a comparison of three strategies for intervention through family members. Journal of Consulting and Clinical Psychology, 1999;67(5):688-697. pubmed.ncbi.nlm.nih.gov/10535235
• Roozen HG, de Waart R, van der Kroft P. Community reinforcement and family training: an effective option to engage treatment-resistant substance-abusing individuals in treatment. Addiction, 2010;105(10):1729-1738. ncbi.nlm.nih.gov/books/NBK79038
• Canadian Centre on Substance Use and Addiction. Community Reinforcement and Family Training, evidence summary, 2017. ccsa.ca

Families researching treatment run into the same number within about ten minutes: something in the region of eighty percent success. It comes up on program websites, in admissions calls, and in conversations with people who mean well. The number is real. It comes from a specific study of a specific population under specific conditions, and almost nobody who repeats it says which. Here is what the research actually supports, with the caveats attached, because a statistic without its conditions is not evidence. It is marketing.

Where the eighty percent comes from

The source is a 2008 cohort study in The BMJ by McLellan, Skipper, Campbell and DuPont, following 904 physicians enrolled in state physician health programs. Of the 802 with known outcomes, 647 (80.7 percent) completed treatment and resumed practice under supervision and monitoring. Of those, 523 (81 percent) had no substance use detected across the full five years of monitoring. At five years, 631 of 802 (78.7 percent) were licensed and working (McLellan et al., 2008).

Now the conditions, which matter more than the headline.

These were physicians. The intervention was not a thirty-day admission. It was residential or intensive treatment followed by five years of structured monitoring: random testing, mandated mutual-help attendance, workplace supervision, a case manager who did not go away, and a medical license contingent on all of it. It is one of the most intensive continuing-care packages ever studied.

So the honest reading is not that treatment succeeds eighty percent of the time. It is that a five-year monitored continuing-care package, delivered to motivated professionals with a career at stake, produced outcomes far better than the ones most people are quoted. What the study demonstrates is the value of duration and accountability. It is not a general population success rate, and any program that presents it as one is doing something you should notice.

What the twelve-step evidence supports

The other figure that circulates loosely is the effectiveness of twelve-step involvement. The strongest evidence here is the 2020 Cochrane review by Kelly, Abry, Ferri and Humphreys, covering Alcoholics Anonymous and manualized twelve-step facilitation for alcohol use disorder.

The finding: manualized AA/TSF outperformed comparison treatments on continuous abstinence, with a relative risk of 1.21 (95 percent CI 1.03 to 1.42) at twelve-month follow-up, and advantages sustained at longer follow-up points. In Project MATCH, 24 percent of outpatients assigned to twelve-step facilitation were continuously abstinent through the first year, against 15 percent for cognitive behavioural therapy and 14 percent for motivational enhancement therapy (Kelly et al., distilled in Alcohol and Alcoholism, 2020).

That is a genuine, replicated, peer-reviewed advantage, and Cochrane also found AA/TSF typically achieved it at lower cost. It is also, plainly, a 24 percent figure and not an 80 percent one. Both things are true. A family deserves to hear both.

Duration is the variable that keeps showing up

Across otherwise dissimilar studies, the finding that persists is that time in structured engagement predicts outcome. The National Institute on Drug Abuse states it as a core principle of effective treatment: remaining in treatment for an adequate period is critical, most people need at least three months to meaningfully reduce or stop use, and the best outcomes come with longer durations (NIDA, Principles of Drug Addiction Treatment).

Duration is also where stability appears to consolidate. Dennis, Foss and Scott followed 1,162 people entering treatment over eight years with better than 94 percent follow-up retention, and found that the probability of sustaining abstinence in the coming year rises substantially with the length of abstinence already achieved (Evaluation Review, 2007). Building on that work and on Vaillant’s long-term alcoholism cohorts, a 2015 paper in the Journal of Substance Abuse Treatment argued for five years as the standard for assessing whether treatment worked, on the grounds that by that point relapse risk approaches that of the general population (Five-Year Recovery).

Note what that implies about a thirty-day program measured at discharge. It is measuring the first two percent of the relevant window.

The number families rarely hear

One more, because the data cuts both ways and this side of it gets quoted far less often. A 2017 national survey published in Drug and Alcohol Dependence estimated that 9.1 percent of US adults, roughly 22.35 million people, have resolved a significant substance use problem (Kelly, Bergman, Hoeppner, Vilsaint and White, 2017).

Resolution is common. It is also, in the data, rarely fast and rarely linear.

What we take from all of this

We are not a treatment program and we do not make outcome claims. What the literature changes about how we work is straightforward.

It says the length and continuity of support matter more than the intensity of any single episode. It says the accountability structure, someone whose attention does not end at a discharge date, is doing a large share of the work in the studies with the best results. It says that a plan measured in years is a different object from a plan measured in weeks, and that families should ask what happens in month four before they ask about the amenities.

When somebody quotes you a success rate, ask three questions. Which study. Which population. Measured at what point after discharge. If the answers are not immediate, you have learned something useful anyway.

For what an engagement looks like in practice, see how we help. If you are weighing programs right now, Choosing a treatment center covers the questions worth asking on the tour.

Sources

• McLellan AT, Skipper GS, Campbell M, DuPont RL. Five year outcomes in a cohort study of physicians treated for substance use disorders in the United States. The BMJ, 2008;337:a2038. pmc.ncbi.nlm.nih.gov/articles/PMC2590904
• Kelly JF, Abry A, Ferri M, Humphreys K. Alcoholics Anonymous and 12-Step Facilitation Treatments for Alcohol Use Disorder: A Distillation of a 2020 Cochrane Review. Alcohol and Alcoholism, 2020;55(6):641-651. academic.oup.com/alcalc/article/55/6/641/5867689
• National Institute on Drug Abuse. Principles of Drug Addiction Treatment: A Research-Based Guide, Third Edition. podat-3rdEd-508.pdf
• Dennis ML, Foss MA, Scott CK. An eight-year perspective on the relationship between the duration of abstinence and other aspects of recovery. Evaluation Review, 2007;31(6):585-612. pubmed.ncbi.nlm.nih.gov/17986709
• Five-Year Recovery: A New Standard for Assessing Effectiveness of Substance Use Disorder Treatment. Journal of Substance Abuse Treatment, 2015. jsatjournal.com
• Kelly JF, Bergman BG, Hoeppner BB, Vilsaint CL, White WL. Prevalence and pathways of recovery from drug and alcohol problems in the United States population. Drug and Alcohol Dependence, 2017;181:162-169. recoveryanswers.org

If you are a parent trying to work out where a good plan goes wrong, here is the short answer: it usually goes wrong in the days between two levels of care, not inside either one of them. Treatment settings are staffed, structured, and accountable. The stretch between discharge and the first outpatient appointment is none of those things, and it is measurably where people are lost.

The gap is measured, and it is wide

The federal government tracks this. The quality measure stewarded by SAMHSA for continuity of care after inpatient or residential substance use treatment reports 2021 national medians of 30.8 percent of discharges followed by a substance use treatment service within seven days, and 37.9 percent within fourteen days (CBE #3453, Partnership for Quality Measurement).

Read that slowly. Roughly two out of three adults leaving residential or inpatient care have no recorded treatment contact in the following week. Not a poor-quality contact. No contact.

That number is not a verdict on any one program. It is a description of a system in which nobody is assigned the days between. The residential clinical team’s responsibility ends at discharge. The outpatient provider’s responsibility begins at intake. The days in between belong to the family, usually with no notice, no training, and no sleep.

Why those particular days carry so much risk

Tolerance falls during a period of abstinence. Behavior, in a hard week, does not always fall with it. That combination makes the weeks immediately after inpatient treatment the highest-risk window in the whole sequence.

A Norwegian eight-year prospective study of drug users discharged from inpatient treatment found a mortality rate ratio of 15.7 in the first four weeks after discharge compared with the rest of the follow-up period. Every death in that window was an opioid overdose (Ravndal and Amundsen, Drug and Alcohol Dependence, 2010, abstract).

That is one cohort in one country, and the absolute numbers behind the ratio are small. We cite it because the direction of the finding has been replicated repeatedly across settings that involve a period of enforced abstinence followed by an unsupervised return. The shape of the risk is consistent even where the size of it varies.

What actually helps, according to the research

The evidence on continuing care is real but it is specific, and the specificity matters more than the headline.

James McKay’s review for the NIAAA journal Alcohol Research: Current Reviews summarizes nineteen randomized trials showing a small but significant benefit for continuing care, with effect sizes of g = 0.19 at the end of the intervention and g = 0.27 at follow-up. The useful part is what separated the trials that worked from the ones that did not: longer planned duration, at least twelve months, and active effort to keep people engaged rather than a standing offer of an appointment (McKay, 2021).

So the variable is not whether aftercare was recommended on the discharge paperwork. It is whether somebody kept showing up to make it happen, for long enough that the habit outlasted the motivation.

What a family can put in place before discharge

None of this requires clinical training. It requires that the days be assigned to a person by name.

• Name who owns the first fourteen days. Not the family collectively. One person, with a phone number, who knows the whole story and does not need it re-explained.
• Book the first outpatient appointment before the discharge date, not after. If the first available slot is nineteen days out, that is your risk window, and you now know its exact length.
• Map the unstructured hours. The first Saturday, the first evening alone, the first paycheck, the first family argument. Put something in each of them on purpose.
• Write down who gets called, in what order, when something feels wrong. Before it feels wrong.
• Ask the discharging program for the actual handoff. A named contact at the next provider, a records release signed, a confirmed appointment. Not a printed list of local numbers.
• Ask about overdose risk directly, including whether naloxone should be in the house and who is trained to use it. This is a question for the clinical team, and it is a fair question to ask out loud.

Where we sit

Alongside is a non-clinical service. We do not treat, diagnose, or replace the clinical team. What we do is take ownership of the stretch that the clinical structure does not cover: the transport, the first weeks home, the standing contact, the person who is already in the car when the plan meets its first real test.

People rarely lose their footing in treatment. They lose it in the gaps between. If you are looking at a discharge date right now, the useful work is not choosing the next program. It is deciding, this week, who owns the days in between and what they are going to do with them.

If you want to think it through with a person, speak with an advisor. If you are earlier than that, Ninety days out covers the same window in more practical detail, and how we help sets out what we actually do.

Sources

• Partnership for Quality Measurement, CBE #3453, Continuity of Care After Inpatient or Residential Treatment for Substance Use Disorder (steward: SAMHSA), 2021 national medians. p4qm.org/measures/3453
• Ravndal E, Amundsen EJ. Mortality among drug users after discharge from inpatient treatment: an 8-year prospective study. Drug and Alcohol Dependence, 2010;108(1-2):65-69. pubmed.ncbi.nlm.nih.gov/20022184
• McKay JR. Impact of continuing care on recovery from substance use disorder. Alcohol Research: Current Reviews, 2021;41(1):01. arcr.niaaa.nih.gov

A family called us this week, relieved: their son had a discharge date. We were glad too, and we said the thing we always say. The discharge date is not the finish line; it is the starting gun. The thirty days after residential care carry more relapse risk than almost any other window, and most of that risk lives in unstructured hours: the first Saturday with no schedule, the first paycheck, the first argument.

What helps is not more willpower. It is a week that has a shape: a person to call before the slip instead of after, standing appointments that outlast motivation, and a family that knows what to say when the honeymoon wears off. Build the reentry plan before the last day of treatment, not after it.

The calls we get are rarely about the loud weeks. Crises announce themselves; everyone rallies. The dangerous stretch is the quiet one: ninety days in, things look stable, and everyone quietly stops paying attention at the same time.

If your loved one is in a quiet stretch, resist the urge to declare victory and change the subject. Keep the standing check-in, keep the meeting schedule, keep saying the recovery words out loud at dinner. Stability is not the absence of the illness; it is the presence of structure. The structure should not leave just because the crisis did.

Most of what a recovery coach does sounds unremarkable when you list it: a morning text, a ride planned in advance, a hard conversation rehearsed before it happens. None of it looks like a breakthrough. All of it is the difference between a bad afternoon and a bad month.

The measure we care about is simple: when something goes sideways, who does the client call first? When the answer is a person who knows the whole story and picks up, the sideways moments stay small. That is the entire job, and it is enough.