Don't do it tonight, and don't do it while they're using. Pick a flat, sober moment, open with something you saw rather than something they are, ask for one specific thing, and expect to do this several times before anything moves. The approach with the best trial evidence behind it is called CRAFT, and you can learn it.
Search this at 2am and you get private clinics selling beds. None mention CRAFT.
First, the thing nobody says to you
You're probably furious, and that isn't a character flaw. Jim Orford's research describes the ordinary family response as being "worried, preoccupied with thinking about the relative and the problem, feeling nervous and panicky, irritable and quick tempered, low and miserable, annoyed and resentful." Resentment is on the list.
What actually gets someone into treatment
CRAFT stands for Community Reinforcement and Family Training. It's a structured programme delivered to the family member rather than the person using. The American Psychological Association describes the goal as to "increase family compliance with an intervention for persons with substance abuse in order to increase the rate of engagement" of those people in treatment, through parts including "communication skills training", "treatment entry training" and "safety training".
In plain English: you learn to spot the openings, what to say in them, and how to have a place ready for the moment they say yes.
The trial everyone cites is Miller, Meyers and Tonigan, 1999.
| Approach | What the family member does | Entry | |---|---|---| | CRAFT | Communication skills, reinforce sober behaviour, have a place ready | Around 64% | | Al-Anon facilitation | Step back, focus on your own recovery | Much lower | | Johnson intervention | Gather the family, confront, present consequences | Much lower |
The published wording: "The CRAFT approach was more effective in engaging initially unmotivated problem drinkers in treatment (64%) as compared with the more commonly practiced Al-Anon (13%) and Johnson interventions (30%)." Treat the two lower numbers as one result. Different reports of the trial split them differently.
Where CRAFT stops working
The 2020 review by Archer and colleagues, in the journal Addiction, pooled 14 studies.
**It didn't work for gambling.** "CRAFT was more than twice as effective when the IP had a substance addiction (RR = 2.35, 95% CI = 1.77-3.12) ... it was no more effective than controls when the IP had a gambling addiction (RR = 0.98, 95% CI = 0.55-1.75)." If the person you love has a gambling problem, this evidence does not transfer.
**The evidence base is thin.** "Quality ratings across the 14 included studies tended to be low: two studies received a 'moderate' rating, and the remaining 12 a 'weak' rating." That doesn't make CRAFT wrong. It makes it the best-supported option in a field where support is modest.
**The book alone doesn't do it.** Self-directed workbooks produced treatment entry rates of 13.3 to 40%. A trained practitioner produced up to 86%. That difference is most of the effect.
The intervention off the television
Here's the finding nobody reports. Families call it off. From the same trial: "most CSOs decide not to go through with the family confrontation (70% in this study) and that among those who do, most (75%) succeed in getting the drinker into treatment." If you've been dreading the big confrontation and quietly suspecting you haven't got it in you, you're the majority.
On confrontation generally, William White and William Miller reviewed the field: "Four decades of research have failed to yield a single clinical trial showing efficacy of confrontational counseling, whereas a number have documented harmful effects, particularly for more vulnerable populations." They name the belief underneath it: "if you can just make people feel bad enough, they will change."
Shame is the fuel of most compulsions. Adding more rarely turns out to be the missing ingredient.
The word "enabling"
Someone will tell you you're enabling, usually someone who isn't in the house. The term comes from the 1970s family-disease model that produced "codependency", and White and Miller are direct: "There never has been a scientific basis for believing that people with substance use disorders, let alone their family members, possess a unique personality or character disorder."
Use a test rather than a label. For any act of help, ask what it actually did. Paying the rent so a child keeps their bed does one thing. Paying a dealer does another. [The enabling question, honestly](/guides/supporters/the-enabling-question-honestly) takes the harder cases.
Two goals, and the content that muddles them
Twelve-step fellowships teach you to detach. CRAFT teaches you to engage. Each camp answers a question the other isn't asking.
For getting someone into treatment, engagement wins the trials. For your own wellbeing, the picture points elsewhere. In a study of coping strategies used by partners, 90% rated "put yourself first, took care of yourself" as at least a little helpful, and 89% said the same of "pursued your own interests". At the bottom sat arguing about the addiction (47%) and getting moody or emotional (45%). That measures perceived helpfulness rather than randomised outcome, so hold it loosely.
You can do a CRAFT course and go to Al-Anon in the same week. [Al-Anon, SMART Family and Friends or the 5-Step Method](/guides/supporters/al-anon-smart-family-and-friends-or-the-5-step-method) compares them.
The conversation itself
- **Pick the time deliberately.** Sober, unhurried, not late, not in front of the children.
- **Decide your one ask first.** "Come to the GP with me on Thursday." Four asks becomes a row about the fourth.
- **Open with what you saw.** "You were asleep on the sofa when the kids came in." A fact has nowhere to go. "You've got a problem" arrives with an argument attached.
- **Say the effect on you, once.** "I was frightened." Not fifteen minutes of the archive.
- **Make the ask, then stop talking.** Let the silence sit longer than is comfortable.
- **Have the practical answer ready.** The number, the appointment, who has the kids. "I'll think about it" often means "I don't know how".
- **When they get angry, don't match it.** "I'm not going to argue. I'll leave it with you." Then leave it.
- **Say it again next week.** Repetition without escalation is the technique.
If it goes badly, that's information rather than a verdict. Note which sentence turned it, and cut that one next time.
Where to actually go
- **SMART Recovery UK's Family and Friends Programme**, which "helps people affected by a loved one's drug or alcohol addiction focus on their own wellbeing", covering communication, boundaries, safety and enabling. Meetings are free and need "no booking or sign-up required". It's recognisably CRAFT-shaped, which is our observation rather than SMART's claim.
- **[We Are With You](https://www.wearewithyou.org.uk/what-we-do/drug-and-alcohol-services-for-adults/support-for-family-and-friends)** runs "specialist support to the family members and friends of those who use substances", including CRAFT, across 80+ free services in England and Scotland.
- **Al-Anon UK, 0800 0086 811**, 10am to 10pm, every day of the year.
- **Families Anonymous UK, 020 7498 4680**, daily 1pm to 4pm and 6pm to 9pm, covering drug use.
Renovyn's [supporter side](/for-supporters) is for the bit between conversations, a record of what you tried. It isn't a course and it isn't treatment.
Common questions
**Can you force someone into rehab in the UK, or get them sectioned for drinking?** We can't give a legal answer here, and the sites promising one are selling beds. The route with the best trial evidence, when someone isn't asking for help, is something you learn and do. If you're worried about their immediate safety, call 999.
**They don't think they have a problem. Is there any point talking to them?** Yes, and that's precisely who the CRAFT trial recruited. Everyone in it was refusing treatment, and around 64% went anyway. The change came from the family changing how they responded, rather than from winning the argument about whether there was a problem.
**What if raising it makes them worse?** Confrontational approaches have documented harms, which is why the script above avoids shaming, ultimatums and audiences. If raising anything frightens you because of how they react, safety comes first. [What is actually yours to carry](/guides/supporters/what-is-actually-yours-to-carry) starts there.
Sources
- [Miller, Meyers and Tonigan, 1999](https://pubmed.ncbi.nlm.nih.gov/10535235/). CRAFT at 64%, and the 70% who declined the confrontation.
- [Archer et al., Addiction, 2020](https://kcmhr.org/pdf/2019_archer_a.pdf). Gambling, quality ratings, workbook versus practitioner.
- [White and Miller, Confrontation in Addiction Treatment](https://cmcffc.org/wp-content/uploads/2022/06/Confrontation-in-Addiction-Treatment.pdf). Four decades without supporting evidence.
- [SMART Recovery UK, Family and Friends](https://smartrecovery.org.uk/family-and-friends-programme). The programme.
