How Do I Stop Drinking?

The short answer

There is no single "right" way to stop drinking — there are several evidence-based paths, and the best one is the one you'll actually stick with. Medications (naltrexone, acamprosate) roughly double the odds of staying off alcohol and are badly under-used; mutual-help programs like AA work about as well as, and for continuous abstinence somewhat better than, professional therapy; and structured therapies (CBT, motivational approaches) help most when they're stacked on top of other support rather than used alone. Most people who resolve a drinking problem do it over time and often after more than one attempt — and a large share do it with little or no formal treatment at all. One critical safety point up front: if you drink heavily every day and get shaky, sweaty, or sick when you stop, quitting suddenly can be medically dangerous — see a clinician before you stop, because you may need supervised detox.

What the evidence actually says

Stopping is common — and rarely a single, clean event. According to PubMed, Deborah Dawson and colleagues' analysis of the U.S. NESARC survey (2005, Addiction) found substantial recovery from alcohol dependence in the general population, with only about a quarter of people having ever received any treatment (DOI: 10.1111/j.1360-0443.2004.00964.x). Recovery is the rule, not the exception — but it usually unfolds over time, and often across several attempts. That's biology and habit, not weakness.

Medication is the most under-used effective tool. A 2014 JAMA systematic review and meta-analysis by Daniel Jonas and colleagues — 122 randomized trials, over 22,000 people — found that two oral medications reliably help. For acamprosate, complete sobriety was achieved by about 1 out of every 12 patients; for naltrexone, about 1 out of 20 eliminated all drinking, while 1 out of 12 to were able to avoid a return to heavy drinking (DOI: 10.1001/jama.2014.3628). Those are strong numbers for behavioral medicine — comparable to many widely prescribed drugs — yet most people who could benefit are never offered them. The COMBINE trial (Raymond Anton and colleagues, 2006, JAMA, over 1,300 patients) confirmed that naltrexone paired with structured medical management improves outcomes, and notably that it can be delivered in ordinary healthcare settings, not just specialty rehab (DOI: 10.1001/jama.295.17.2003).

AA and Twelve-Step Facilitation work — and the best evidence now says so. For decades, "does AA actually work?" was answered mostly with anecdote. According to PubMed, that

changed with the 2020 Cochrane systematic review by John Kelly, Keith Humphreys, and Marica Ferri — 27 studies, over 10,000 participants — which found that manualized Twelve-Step Facilitation (a clinician-delivered way of connecting people to AA) produced higher rates of continuous abstinence than other established treatments like CBT (about 42% vs. 35% at some follow-ups), performed at least as well on other drinking outcomes, and generated meaningful healthcare cost savings (DOI: 10.1002/14651858.CD012880.pub2). Importantly, many of the 12-step participants did not provide complete data because of high drop-out rates, which are common in AA and the 12-step context. Nevertheless, if total abstinence is your goal and the fellowship model fits you, this is a genuinely strong, free, widely available option.

Therapy helps most as an add-on, not a stand-alone. A 2026 meta-analysis by Molly Magill and colleagues in Behaviour Research and Therapy — 52 randomized trials, over 9,400 participants — found that cognitive behavioral therapy for substance use did not outperform other evidence-based treatments head-to-head, but clearly beat usual or minimal care, and was strongest when added on top of existing care (a medium effect on both consumption and psychosocial functioning) (DOI: 10.1016/j.brat.2026.105101). The lesson isn't "CBT doesn't work" — it's that no single modality is magic, and stacking supports beats betting everything on one.

Even a brief conversation moves the needle. A 2018 Cochrane review by Eileen Kaner and colleagues — 69 trials, over 33,000 people — found that brief interventions (a short, structured conversation with a primary-care or ER clinician) reduced drinking by roughly 20 grams of alcohol per week versus no intervention (DOI: 10.1002/14651858.CD004148.pub4). You don't need a residential program to start; sometimes one honest conversation with your doctor is a real first step.

You don't have to be perfect to benefit. Katie Witkiewitz and colleagues (2017, Alcohol Clin Exp Res) showed that even reducing your drinking — not only quitting — produces measurable improvements in consequences and mental health (DOI: 10.1111/acer.13272). If "stop entirely, forever, starting today" feels impossible, cutting down is a legitimate and beneficial direction of travel. (For whether moderation can be your endpoint rather than a step toward abstinence, see the companion Atlas article, Can I moderate my drinking instead of quitting?)

The safety exception, stated plainly. Alcohol is one of the few drugs where withdrawal itself can be life-threatening. If you are physically dependent — daily heavy drinking, and shakes, sweating, nausea, anxiety, or a racing heart when you don't drink — stopping abruptly can cause seizures or delirium tremens. This is a medical emergency risk, not a willpower issue. Talk to a clinician about a supervised taper or detox before you quit cold.

Where the experts disagree

Is abstinence the only real goal, or is reduction legitimate? The AA-rooted, abstinence-first tradition holds that for genuine dependence, "stopping" means stopping completely and permanently. The harm-reduction tradition counters that reduced drinking is a valid, beneficial outcome and a better on-ramp for the many people not ready to quit forever. The evidence supports both being true for different people — reduction produces real benefit (Witkiewitz et al., 2017), and severe dependence responds better to abstinence goals (Dawson et al., 2005). This is covered in depth in the companion article.

Does AA deserve its central place? Critics point to high dropout, the discomfort of the disease/powerlessness language for some people, and selection effects in older research. Kelly's 2020 Cochrane review answered much of the efficacy question — manualized TSF genuinely helps — but "it works on average" doesn't mean "it works for everyone," and the fellowship's spiritual framing is a poor fit for a meaningful minority. Fit matters more than ideology.

Medication vs. "real" recovery. A stubborn cultural belief — sometimes inside recovery communities themselves — treats medication as a crutch or as "not really sober." The pharmacology data (Jonas 2014; Anton 2006) make this position hard to defend on evidence: naltrexone and acamprosate measurably help, and withholding them on principle costs people outcomes. This is one place where the science is well ahead of parts of the culture.

Where nearly everyone agrees: most people take more than one attempt; matching the method to the person beats any one-size-fits-all program; and the biggest failure in the system is not that the wrong method gets used, but that most people are never offered the effective options at all.

My take: stop asking "which program?" and start asking "what's the drink for?"

Most people approach quitting as a logistics problem — which program, which app, which meeting. Those matter. But the reason so many attempts collapse in week three isn't that someone picked the wrong program. It's that they tried to remove the drinking without ever addressing what the drinking was doing for them.

In my framework, a compulsive behavior sits inside a hook cycle: it persists because it reliably delivers something you need — an off-switch for anxiety, a bridge into sleep, a social lubricant, a boundary at the end of a brutal day. My SPARO framework maps that pathway from the

triggering stimulus through to the behavior and its outcome, and the EAT sequence (Emotion → Attention → Thought) names where it usually starts: not with a thought like "I want a drink," but with an emotion the drink has learned to answer.

Here's why that changes how you should stop. If alcohol is currently your main tool for managing anxiety, and you remove it with nothing in its place, you haven't solved anything — you've deleted your coping mechanism and kept the problem. The craving that follows isn't a sign you're a hopeless case. It's the predictable result of an unmet need looking for its usual answer. This is why the durable question isn't just "how do I stop?" but "what will I put in the place of what drinking was doing?"

So my practical sequence is:

  1. Handle safety first (withdrawal risk - above).

  2. Find the function. What does the drink reliably give you? That's the thing you actually have to replace.

  3. Stack your supports instead of betting on one. The evidence is unambiguous that combinations outperform any single tool: medication + a mutual-help or therapy connection + a plan for the underlying need.

  4. Expect iteration. Most people who succeed have a "failed" attempt or two behind them. Those aren't failures; they're the map. The programs are the how. The function is the why it keeps not working — and that's the part almost nobody addresses first.

What to actually do

  • ✅ Assess withdrawal risk before anything else. Daily heavy drinking plus shakes/sweats/nausea when you stop = see a clinician about supervised detox before quitting. Non-negotiable.

  • ✅ Ask a doctor about medication — specifically naltrexone or acamprosate. They're evidence-based, under-prescribed, and don't require specialty rehab to get. This is the highest-leverage, most-overlooked step.

  • ✅ Pick a support structure that actually fits you. AA/Twelve-Step Facilitation (strong evidence, free, everywhere) if the fellowship model resonates; SMART Recovery or other secular options if the spiritual framing doesn't. Fit predicts follow-through.

  • ✅ Add a therapy or coaching layer if you can — on top of, not instead of. CBT and motivational approaches work best stacked with other supports.

  • ✅ Name and replace the function. For one week, note what you felt in the hour before each drink. Then deliberately build a replacement for that specific need — not just a rule against drinking.

  • ✅ If "forever" feels impossible, start by reducing. A brief conversation with your doctor and a concrete cut-down plan are real progress and often the on-ramp to more.

  • ✅ Treat setbacks as data, not verdicts. A slip is information about your triggers, not proof you can't do this. Most successful quits include a few.

FAQ

What is the most effective way to stop drinking? There's no single most-effective method for everyone; the strongest evidence supports combining approaches — medication (naltrexone or acamprosate; Jonas et al., 2014, DOI: 10.1001/jama.2014.3628), a mutual-help or therapy connection (Kelly et al., 2020, DOI: 10.1002/14651858.CD012880.pub2; Magill et al., 2026, DOI: 10.1016/j.brat.2026.105101), and a plan for the need the drinking was meeting.

Can I just stop drinking on my own, cold turkey? Many people do resolve drinking problems with little formal treatment (Dawson et al., 2005, DOI: 10.1111/j.1360-0443.2004.00964.x) — but "cold turkey" is genuinely risky if you're physically dependent. Daily heavy drinking with withdrawal symptoms means abrupt cessation can cause seizures; see a clinician first. Without physical dependence, self-directed change is a reasonable path, ideally with some support.

Do I need rehab to stop drinking? Not necessarily. Effective options include primary-care medication management, mutual-help groups, outpatient therapy, and brief interventions — a short conversation with a clinician alone reduces drinking (Kaner et al., 2018, DOI: 10.1002/14651858.CD004148.pub4). Residential rehab is one option among several, not a prerequisite.

Does AA actually work? Yes, for those who stay with it, on the best current evidence. The 2020 Cochrane review found manualized Twelve-Step Facilitation produced higher continuous-abstinence rates than treatments like CBT, at lower healthcare cost (Kelly et al., 2020, DOI: 10.1002/14651858.CD012880.pub2). It's not the only thing that works, and the fellowship model isn't for everyone — but "AA has no evidence" is out of date.

Is it bad to take medication to stop drinking — isn't that just trading one drug for another? No. Naltrexone and acamprosate aren't intoxicating and aren't addictive; they reduce craving and the reward from alcohol, and the meta-analytic evidence shows they measurably

improve outcomes (Jonas et al., 2014, DOI: 10.1001/jama.2014.3628). The idea that using them isn't "real" recovery isn't supported by the evidence.

How long does it take to stop drinking for good? For most people it's a process, not a single event, and often includes more than one attempt before it holds (Dawson et al., 2005, DOI: 10.1111/j.1360-0443.2004.00964.x). That's normal and not a sign of failure — the people who succeed are usually the ones who keep adjusting after a setback.

References

  1. 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. DOI: 10.1001/jama.2014.3628

  2. Kelly JF, Humphreys K, Ferri M. Alcoholics Anonymous and other 12-step programs for alcohol use disorder. Cochrane Database Syst Rev. 2020;3(3):CD012880. DOI: 10.1002/14651858.CD012880.pub2

  3. Anton RF, O'Malley SS, Ciraulo DA, et al. Combined pharmacotherapies and behavioral interventions for alcohol dependence: the COMBINE study: a randomized controlled trial. JAMA. 2006;295(17):2003–2017. DOI: 10.1001/jama.295.17.2003

  4. Magill M, Nichols LM, Kiluk BD, Alton E, Tartak O, Ray LA. A meta-analysis of cognitive behavioral therapy for substance use disorder: Treatment effects by comparator type and consumption and psychosocial outcomes. Behav Res Ther. 2026;203:105101. DOI: 10.1016/j.brat.2026.105101

  5. Kaner EFS, Beyer FR, Muirhead C, et al. Effectiveness of brief alcohol interventions in primary care populations. Cochrane Database Syst Rev. 2018;2(2):CD004148. DOI: 10.1002/14651858.CD004148.pub4

  6. Dawson DA, Grant BF, Stinson FS, Chou PS, Huang B, Ruan WJ. Recovery from DSM-IV alcohol dependence: United States, 2001–2002. Addiction. 2005;100(3):281–292. DOI: 10.1111/j.1360-0443.2004.00964.x

  7. Witkiewitz K, Hallgren KA, Kranzler HR, et al. Clinical Validation of Reduced Alcohol Consumption After Treatment for Alcohol Dependence Using the World Health Organization Risk Drinking Levels. Alcohol Clin Exp Res. 2017;41(1):179–186. DOI: 10.1111/acer.13272

Limitations stated honestly: Effect sizes in behavioral medicine are modest and vary substantially by population and study design (the CBT and AA reviews both report meaningful heterogeneity). Meta-analytic averages describe groups, not any single reader. Numbers-needed-to-treat depend on the outcome defined and the follow-up window. And none of this replaces an individual clinical assessment — especially where withdrawal risk is present.

Written by Dr. Adi Jaffe, PhD (UCLA), author of The Abstinence Myth and Unhooked. Last medically reviewed: [July-17-2026].