Drinking Less as a Health Question: What the Research Shows

A Health Question, Not a Willpower Question

Cutting back on alcohol is usually framed as a test of discipline, with the person on the receiving end told to want it more. The evidence describes something else.

Alcohol use disorder (AUD) is a diagnosed medical condition. NIAAA, the National Institute on Alcohol Abuse and Alcoholism, classifies it as a brain disorder and notes that it ranges from mild to severe. NIAAA reports that 27.1 million US adults, about 10.3%, had AUD in the past year, based on the 2024 National Survey on Drug Use and Health. Globally, an estimated 400 million people live with alcohol use disorders, according to the World Health Organization (WHO).

Two approaches have real evidence behind them: behavioral support and medication. Both have been studied at scale, and both reach far fewer people than the size of the problem suggests.

The Behavioral Evidence: What Actually Helps People Cut Back

The behavioral evidence begins with a Cochrane review of personalized, digital interventions, synthesizing 41 studies and 42 comparisons with 19,241 participants. In the studies performed, participants who received a digital intervention reported average drinking of 23 grams per week fewer than those receiving no digital intervention or minimal. This amounts to just over 1½ US standard drinks. The same review put the drop in binge-drinking occasions at about one a month.

The review also highlights three techniques associated with the greater reductions, which include behavior substitutions, problem solving, and credible source (advice provided by a trusted and respected person or organization). The averages mask key constraints. This digital versus face-to-face comparison is based on five small studies, and the review is based on poor quality evidence.

No single element of any drinking-reduction app seems to be the “magic bullet”. Five modules were tested in a factorial experiment published in Scientific Reports in 2018 and none of the modules had a significant main effect on the consumption per week. It was the other way around in terms of usage and ratings: the increased self-monitoring and feedback module was used much more.

often, and the one they rated significantly more helpful. The review’s own follow-up breakdown shows the effect was largest early: 43.3 grams per week at two to three months, then 11.5 grams per week beyond three months.

The Medication Evidence: What Randomized Trials Show

The medication evidence runs deeper than behavioral health research usually does. One 2023 review in JAMA combined the results of 118 clinical trials and 20,976 participants.

Oral naltrexone is an opioid antagonist: it blocks opioid receptors and dampens the rewarding effect of alcohol. In that review, the number needed to treat for a return to heavy drinking was 11 (95% CI 5 to 41), and 18 for a return to any drinking (95% CI 4 to 32). In plain terms, 11 people have to be treated for one of them to avoid a return to heavy drinking. That review places oral naltrexone and acamprosate among the first-line pharmacotherapies, with a qualification that matters: they are used in conjunction with psychosocial interventions.

The COMBINE trial followed 1,383 participants who received naltrexone or a placebo alongside medical management. It reported better drinking outcomes in the naltrexone group, and it was published in JAMA in 2006. The FDA label itself is more measured than the enthusiasm around the drug. It states that the mechanism “is not understood,” and it describes naltrexone as a treatment given “as an adjunct to social and psychotherapeutic methods.” The expected effect, the label adds, is “a modest improvement,” and the drug “was not uniformly helpful to all patients.”

The label reports the two arms of a 104-patient, 12-week trial: abstention was 51% on naltrexone against 23% on placebo, and relapse was 31% against 60%. Naltrexone was approved for alcohol dependence in 1994, ten years after its approval for opioid dependence. The label also sets two limits. It carries a warning about liver injury, and it states that naltrexone is contraindicated for people who are currently dependent on opioids or in acute opioid withdrawal. Nausea, headache, fatigue, and dizziness are listed among the common side effects.

The Gap Between Evidence and Uptake

If the evidence is this specific, why do so few people use it? NIAAA’s analysis of national survey data found that only 7.6% of people aged 12 and older with past-year AUD received any alcohol use treatment in the past year. Only 2.5%, about 697,000 people, received medication-assisted treatment. Global figures are lower still. WHO reports that treatment contact ranges from under 1% in some countries to no more than 35% in others. This is a delivery problem rather than a discovery problem. Medications approved for three decades, and behavioral programs tested in trials covering more than 19,000 people, are not reaching most of the people they were studied for.

Why This Is a Health Question: Risk and Prevention

It’s a health case for less because of risk research. NIAAA’s overview of alcohol effects focuses not only on the liver, but on the whole body—including the brain, gut, pancreas, lungs, cardiovascular system, and immune system—effects. In addition, it states that current research suggests that there are health risks, even at low levels of alcohol consumption, regardless of beverage. The US Surgeon General’s 2024 recommendation on alcohol and cancer risk extends beyond this. It refers to evidence for a causal relationship between alcohol use and seven types of cancer: breast cancer in women, colorectal, esophageal (food pipe), voice box (larynx), liver, mouth and throat. The advisory estimates that worldwide, 741,300 cancer cases in 2020 were linked with drinking alcohol. The death toll is even higher. In the USA in 2020-2021, alcohol-related deaths accounted for approximately 178,000 fatalities annually, which has increased by 29% compared to 2016-2017. WHO reports that alcohol was responsible for 2.6 million deaths in the world in 2019, representing 4.7% of all deaths.

Common Questions

Does cutting back require treatment?

No single pathway is required. The evidence reviewed here covers self-guided behavioral techniques and, where it fits, medication. Nothing in the research reviewed here says one route is right for everyone.

Is medication a replacement for counseling?

No. The FDA label describes naltrexone as an adjunct to social and psychotherapeutic methods, and the JAMA review conditions its first-line recommendation on pairing medication with psychosocial interventions. The Substance Abuse and Mental Health Services Administration (SAMHSA) places medication inside a comprehensive treatment plan.

How long do behavior changes hold?

Reported reductions were largest in the first few months and smaller beyond three months, which points toward ongoing support instead of one concentrated effort. That is a pattern in group averages, not a forecast for any individual.

What the Evidence Supports

Together, the three bodies of evidence show the same pattern. Behavioral techniques that a person can learn produce modest average reductions in drinking. Medication does the same, with trial data that is unusually specific about how large the benefit is. The FDA label and the JAMA review agree that the drug works alongside other treatment.

Among the programs built on that research, Sunnyside combines drink tracking and coaching with behavior-change techniques drawn from the same studies. Nothing in this evidence promises a given result for a given person. The average effects are real, and they are modest.

Conclusion

Drinking less belongs in a health section, next to blood pressure and cancer screening, because the evidence exists on both sides of it. Behavioral support works on average, and medication works on average. Most people with AUD reach neither, and that gap is the practical problem. Decisions about medication should be made with a qualified clinician.