Skip to content
MyFreud

Addiction: How It Works, and What Treatment Looks Like

A 2026 guide to addiction: the science behind substance and behavioural addictions, why clinicians treat it as a health condition, and which treatments work.

9 min read

Pop-art illustration of a man in glasses looking at his own reflection in a mirror.

Key takeaways

  • Addiction is a chronic health condition rather than a failure of willpower. It produces lasting changes in the brain circuits governing reward, motivation and self-control, which is why stopping is harder than deciding to.
  • The brain disease model describes three stages, each involving different brain systems. Intoxication runs on dopamine surges far larger than food or company produce, and that is what makes a substance reinforcing so quickly.
  • DSM-5-TR diagnoses a substance use disorder on at least two of eleven criteria met within twelve months, graded mild, moderate or severe. The criteria group around impaired control, social harm, risky use, and tolerance or withdrawal.
  • For opioid use disorder, medication-assisted treatment has some of the strongest evidence in addiction medicine, with two Cochrane reviews behind it. Medication and psychosocial support do better together than either alone.
  • Recovery rarely looks like one decision that holds. For many people it unfolds over years, with stability, partial return to use, and renewed effort, so a relapse is a point in the process rather than the end of it.

Addiction is a chronic health condition characterised by the compulsive use of a substance or engagement in a behaviour, continued despite serious harm and a genuine desire to stop. It is not a character flaw or a failure of willpower. That framing, however intuitive it might feel, does not fit the evidence: research consistently shows that addiction produces lasting changes in the brain circuits that govern reward, motivation, and self-control, changes that make stopping far harder than simply deciding to. The global burden is substantial. The GBD 2016 analysis by Degenhardt et al. (2018) estimated that opioid dependence affected approximately 26.8 million people worldwide and alcohol use disorders approximately 100.4 million, contributing around 99.2 million disability-adjusted life years from alcohol alone. [degenhardt-2018-gbd] Treating addiction as a health condition is both scientifically accurate and the approach most likely to lead people toward effective help.

What addiction is, and what happens in the brain

The dominant scientific framework for understanding addiction is the brain disease model, articulated most fully in a 2016 review in the New England Journal of Medicine by Volkow, Koob, and McLellan. [volkow-2016-brain-disease] Their account describes addiction as progressing through three interconnected stages, each involving distinct brain systems.

The same questions come up about gaming, usually measured the wrong way. Is my gaming a problem? covers why hours are the weakest indicator and what the criteria actually turn on.

During the intoxication stage, the brain’s reward circuits, particularly the dopamine pathways running through the nucleus accumbens, are strongly activated. Drugs and alcohol produce dopamine surges far larger than those triggered by natural rewards such as food or social connection, which is why they can become highly reinforcing very quickly.

As use becomes regular, the withdrawal stage takes hold. The reward system becomes less sensitive, so the same amount of a substance produces less pleasure, while stress-related circuits become more active. The person begins using not primarily to feel good but to escape feeling bad: to relieve anxiety, dysphoria, and physical discomfort. This negative reinforcement cycle is one of the most powerful drivers of continued use.

The preoccupation stage involves changes in the prefrontal cortex, the part of the brain responsible for planning, self-regulation, and resisting impulses. Its function is reduced, impairing the ability to weigh long-term consequences against short-term cravings. This explains why a person can want, sincerely and simultaneously, to stop using and to use again within hours.

The brain disease model, while widely accepted in clinical medicine and public health, is not without academic critics, who argue it underweights social, contextual, and volitional factors. The most defensible reading is that these neurobiological changes are real and clinically significant, but they do not fully determine behaviour: context, relationship, and meaning matter too. What is not in serious dispute is that addiction is treatable.

Substance and behavioural addictions: where the line falls

Substance use disorders cover alcohol, opioids, cannabis, stimulants such as cocaine and amphetamines, and a range of other drugs. The DSM-5-TR (the 2022 text revision of the Diagnostic and Statistical Manual) uses the umbrella term substance use disorder, with severity specified as mild, moderate, or severe based on the number of criteria met. The ICD-11, the international classification system used by the World Health Organization, uses dependence and harmful use as its main categories.

Gambling is the clearest example, and the ordinary end of it is worth understanding too: the psychology of the lottery covers why a ticket feels good even when you can recite the odds, and where routine play tips into harm.

Behavioural addictions are a newer and more contested category. Currently, the two formal diagnoses recognised in mainstream systems are gambling disorder, which appears in both the DSM-5-TR and ICD-11, and gaming disorder, which is included in the ICD-11 under impulse control disorders (effective January 2022) but not yet in DSM-5-TR. Both meet the formal threshold because the evidence supports a pattern of escalating preoccupation, loss of control, and continued behaviour despite significant harm.

Other patterns, including compulsive sexual behaviour, heavy social media use, and problematic food consumption, are sometimes discussed under the addiction label. ICD-11 does include compulsive sexual behaviour disorder, but explicitly classifies it under impulse control disorders rather than addictive disorders. Clinicians remain cautious about extending the addiction frame too broadly, partly because the neurobiological evidence is thinner and partly because the label can pathologise intense engagement that is not causing harm.

Gambling disorder is more common than often assumed. A 2024 meta-analysis by Tran et al. in Lancet Public Health, drawing on 380 representative samples across 68 countries, estimated that 1.41 per cent of adults engage in problematic gambling globally, with substantially higher rates among those who gamble online. [tran-2024-gambling] For gaming disorder, a 2024 review by Saunders et al. in Current Opinion in Psychiatry concluded that between 3 and 6 per cent of the general population meet criteria by ICD-11 or DSM-5 measures, though the authors cautioned that many estimates come from non-representative samples. [saunders-2024-gaming]

How addiction is diagnosed and why it is not a moral failing

Under the DSM-5-TR, a substance use disorder is diagnosed when a person’s pattern of use leads to clinically significant impairment or distress and meets at least two of eleven criteria within a twelve-month period. These criteria group around four themes: impaired control over use (taking more than intended, failed attempts to cut down, spending significant time obtaining or recovering from the substance, craving); continued use despite social problems it is causing; risky use; and pharmacological markers where applicable, meaning tolerance and withdrawal.

The severity of the disorder, mild at two to three criteria, moderate at four to five, and severe at six or more, matters clinically because it guides treatment intensity. Someone with a mild alcohol use disorder and strong social support may benefit from brief counselling and monitoring; someone with severe opioid use disorder and co-occurring depression will typically need a more comprehensive and longer-term programme.

The finding by McLellan et al. (2000), published in JAMA, that addiction shares key features with other recognised chronic conditions, including comparable relapse rates, similar heritability, and the same requirement for ongoing rather than one-time treatment, helped reshape how clinicians and policymakers understand the condition. [mclellan-2000-chronic] Just as someone with asthma or type 2 diabetes is not blamed for needing long-term management, a person with a severe alcohol or opioid use disorder is not failing by needing sustained support. The appropriate response to relapse is to adjust the treatment plan, not to withdraw help.

Evidence-based treatment: medication, therapy, and support

For opioid use disorder, medication-assisted treatment (MAT) is one of the most robustly evidenced interventions in all of addiction medicine. Two Cochrane systematic reviews by Mattick et al. summarise the evidence. Their 2009 review found that methadone maintenance was statistically significantly more effective than non-pharmacological approaches in retaining people in treatment and suppressing heroin use. [mattick-2009-methadone] Their 2014 review found that buprenorphine at high doses showed comparable efficacy to methadone, with flexible dosing slightly less effective in terms of retention but still substantially better than placebo. [mattick-2014-buprenorphine] Both medications work by stabilising opioid receptor activity, reducing craving and withdrawal, and allowing people to re-engage with everyday life.

For alcohol use disorder, both naltrexone and acamprosate have well-established evidence bases. A 2023 systematic review by McPheeters et al. for the Agency for Healthcare Research and Quality found moderate strength of evidence that oral naltrexone at 50 mg reduced return to any drinking, return to heavy drinking, and the percentage of heavy drinking days, while acamprosate showed moderate strength of evidence for reducing return to any drinking. [mcpheeters-2023-aud-pharma] Both are typically used alongside psychological support rather than as standalone treatments.

Psychological therapies with good evidence across substance and behavioural addictions include cognitive-behavioural therapy (CBT), motivational enhancement therapy, and contingency management. Contingency management, in which people receive tangible rewards, typically vouchers or prize draws, contingent on verified abstinence or treatment engagement, has a particularly strong evidence base. Prendergast et al. (2006) conducted a meta-analysis of 47 comparisons and found an overall effect size of d = 0.42, with higher effect sizes for opioid and cocaine use (d = 0.65 and 0.66 respectively). [prendergast-2006-cm] Despite this evidence, contingency management remains underused in clinical services, partly because of persistent philosophical objections to the incentive model.

Mutual aid and peer support, including twelve-step programmes, SMART Recovery, and similar peer-led groups, are often valuable alongside formal treatment, though they function as support rather than as treatment in the clinical sense. The social connection they provide can be powerful, particularly for people whose relationships have been damaged by addiction.

Is it use, or is it dependence?

Tick anything true of the last year. This is a reflection prompt rather than a test, and it produces no diagnosis.

0 of 6 ticked

No screener on this site covers substance use, and we would rather say so than point you at one measuring something else. The hub lists the seven we do have, and our guide to habit versus addiction covers the distinction the checklist above is circling.

Recovery, relapse, and where to find help

Recovery is real, but it rarely looks like a single decision that holds permanently. For many people it is a process that unfolds over years, with periods of stability, partial return to use, and renewed effort. This is not unusual: McLellan et al. (2000) documented that relapse rates for addiction are broadly comparable to those for other chronic conditions like hypertension and asthma, where treatment adherence fluctuates and maintenance is ongoing.

The implication is that relapse is a clinical signal, not a verdict. It indicates that the current treatment plan needs review: perhaps the medication dose is insufficient, the psychosocial support is too thin, or there are unaddressed co-occurring conditions such as depression, anxiety, or trauma. Treating relapse as failure, and withdrawing care accordingly, is one of the factors that makes addiction harder to manage than it needs to be.

People vary considerably in how they define and experience recovery. For some, full abstinence is the goal and the outcome; for others, harm reduction, meaning reducing the harms associated with use without necessarily eliminating it, is a realistic and meaningful aim. Effective services take both seriously.

One thing worth settling before that first appointment is how bad things have to get before it counts. The usual answer, that you have to reach a low point first, does not survive contact with the evidence: brief conversations in primary care reduce drinking in people who were not seeking help at all. Do you have to hit rock bottom to recover goes through what the research actually supports.

If you or someone close to you is affected, a doctor is usually the best first contact: they can assess what support is appropriate, initiate medication if indicated, and refer on to specialist addiction services. In acute situations, crisis lines and emergency services are the right starting point. Recovery is not a straight line, but with appropriate support, it is achievable.

Recent research we have covered

Single studies rarely settle anything on their own, and most of the ones below are small. They are worth reading for what they narrow down.

Frequently asked questions

Is addiction a disease or a choice?

This is one of the most persistent and politically charged questions in addiction science. The short answer is that the evidence supports treating it primarily as a health condition rather than a moral failing, without dismissing the role of choice entirely. The brain disease model, developed by researchers including Volkow, Koob, and McLellan, shows that long-term substance use produces measurable changes in the reward, stress, and executive function circuits of the brain. These changes make it significantly harder to stop, even when the person strongly wants to. At the same time, social context, early life adversity, access to support, and personal meaning all influence how addiction develops and whether recovery happens. The most accurate position is that addiction involves both altered brain function and the full complexity of a person's life, and that effective treatment must address both.

What is the difference between substance and behavioural addiction?

Substance addictions involve physical dependence on a drug or alcohol, with tolerance (needing more to get the same effect) and withdrawal (physical and psychological symptoms when use stops) as common features. Behavioural addictions involve compulsive engagement in an activity, currently gambling disorder and gaming disorder are the two formal diagnoses in major classification systems. Despite this surface difference, both types activate overlapping reward circuits, share patterns of escalating preoccupation and loss of control, and respond to some of the same psychological treatments. The main practical difference is that medication-assisted treatment exists for several substance use disorders but not yet for behavioural addictions.

Is gaming or gambling really an addiction?

Gambling disorder is a well-established diagnosis in both DSM-5-TR and ICD-11, with a substantial evidence base. Gaming disorder was added to the ICD-11 in 2022. Both meet the core clinical criteria: escalating preoccupation, loss of control over the behaviour, continued engagement despite significant harm to relationships, work, or mental health, and distress when the person tries to cut down. Tran et al. (2024) estimated that 1.41 per cent of adults globally engage in problematic gambling, a figure that translates to tens of millions of people. Saunders et al. (2024) found rates of gaming disorder between 3 and 6 per cent in various general population studies, though methodological variability means these estimates should be interpreted with some caution. The key clinical question is not whether gaming or gambling can become addictive in principle, but whether a specific individual's engagement meets the criteria for a disorder.

What treatments actually work for addiction?

The treatments with the strongest evidence depend partly on the type of addiction. For opioid use disorder, methadone and buprenorphine maintenance therapy have the most robust evidence, with Cochrane reviews showing significantly better retention and reduced illicit opioid use compared to no medication. For alcohol use disorder, naltrexone and acamprosate both have moderate evidence supporting their use. Across substance and behavioural addictions, cognitive-behavioural therapy and contingency management have consistent evidence bases. Contingency management in particular shows effect sizes around d = 0.42 to 0.66 depending on the substance, making it one of the more powerful behavioural interventions available. Long-term mutual aid participation, while less formally studied, is associated with sustained recovery for many people.

Does relapse mean treatment has failed?

No. Relapse is common across chronic health conditions, including addiction, and its occurrence does not indicate that treatment has been ineffective or that recovery is impossible. McLellan et al. (2000) found that relapse rates for substance use disorders are broadly similar to those for other chronic conditions like asthma and hypertension, where lapses in adherence or symptom return are routine and managed clinically rather than treated as treatment failure. The appropriate response to relapse is to review what changed, adjust the treatment plan, and continue. Withdrawing care or support after relapse makes outcomes worse, not better.

Can someone recover from addiction fully?

Yes, and many people do. Recovery can mean different things to different people, ranging from full abstinence to a stable, harm-reduced relationship with a substance, to a life in which addiction no longer dominates one's relationships, work, or sense of self. Recovery is often non-linear, with setbacks along the way, but long-term population data consistently show that a substantial proportion of people with severe substance use disorders achieve durable recovery, particularly with access to appropriate treatment and support. The prognosis is better than the condition's reputation suggests.

References

  1. 1.Volkow ND, Koob GF, McLellan AT ( 2016). Neurobiologic Advances from the Brain Disease Model of Addiction. N Engl J Med. pubmed.ncbi.nlm.nih.gov . doi:10.1056/NEJMra1511480
  2. 2.GBD 2016 Alcohol and Drug Use Collaborators (Degenhardt L et al.) ( 2018). The global burden of disease attributable to alcohol and drug use in 195 countries and territories, 1990-2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet Psychiatry. pubmed.ncbi.nlm.nih.gov . doi:10.1016/S2215-0366(18)30337-7
  3. 3.McLellan AT, Lewis DC, O'Brien CP, Kleber HD ( 2000). Drug dependence, a chronic medical illness: implications for treatment, insurance, and outcomes evaluation. JAMA. pubmed.ncbi.nlm.nih.gov . doi:10.1001/jama.284.13.1689
  4. 4.Mattick RP, Breen C, Kimber J, Davoli M ( 2009). Methadone maintenance therapy versus no opioid replacement therapy for opioid dependence. Cochrane Database of Systematic Reviews. pubmed.ncbi.nlm.nih.gov . doi:10.1002/14651858.CD002209.pub2
  5. 5.Mattick RP, Breen C, Kimber J, Davoli M ( 2014). Buprenorphine maintenance versus placebo or methadone maintenance for opioid dependence. Cochrane Database of Systematic Reviews. pubmed.ncbi.nlm.nih.gov . doi:10.1002/14651858.CD002207.pub4
  6. 6.McPheeters M, O'Connor EA, Riley S et al. ( 2023). Pharmacotherapy for Adults With Alcohol Use Disorder in Outpatient Settings: Systematic Review. Agency for Healthcare Research and Quality. pubmed.ncbi.nlm.nih.gov . doi:10.23970/AHRQEPCCER262
  7. 7.Prendergast M, Podus D, Finney J, Greenwell L, Roll J ( 2006). Contingency management for treatment of substance use disorders: a meta-analysis. Addiction. pubmed.ncbi.nlm.nih.gov . doi:10.1111/j.1360-0443.2006.01581.x
  8. 8.Tran LT, Wardle H, Colledge-Frisby S et al. ( 2024). The prevalence of gambling and problematic gambling: a systematic review and meta-analysis. Lancet Public Health. pubmed.ncbi.nlm.nih.gov . doi:10.1016/S2468-2667(24)00126-9
  9. 9.Saunders JB, Chan G, Leung J, Stjepanovic D, Connor JP ( 2024). The nature and characteristics of problem gaming, with a focus on ICD-11 diagnoses. Current Opinion in Psychiatry. pubmed.ncbi.nlm.nih.gov . doi:10.1097/YCO.0000000000000949

Explore the addiction cluster

How to Help an Alcoholic: What the Evidence Says

Confrontation and waiting for rock bottom are the best-known approaches and neither has the best evidence. What a randomised trial found actually works.

Roblox Parental Controls and What They Miss

Roblox is a platform, not a game, so hours per day answers nothing. What the parental controls can set, and the three questions that predict a problem.

Affection Addiction: What Is Actually Happening

Affection addiction is not a diagnosis in any classification system. The pull behind the word is real, and attachment and reinforcement explain it better.

Brad Pitt’s Sobriety: What Recovery Looks Like

Brad Pitt has spoken about years of sobriety. What long-term recovery actually involves, why the risk curve flattens, and what the milestone talk gets wrong.

GTA 6 on Netflix: Is My Gaming a Problem?

GTA 6 will produce a lot of very long sessions. Gaming disorder is real and rare, and hours are the weakest indicator. What the criteria actually turn on.

The Psychology of the Lottery: Why a Ticket Feels Good

Why buying a lottery ticket feels good even though you know the odds, what a near miss does to the brain, and where ordinary play tips into harm.

Relapse: Why One Slip Turns Into a Return

What decides whether a lapse becomes a full relapse is rarely the substance itself. It is what you conclude about yourself in the hour or so afterwards.

Dopamine Detox: What It Gets Wrong, and What Works

You cannot detox from a neurotransmitter your brain makes on purpose. The underlying advice is sound, the mechanism is invented, and the difference matters.

Habit vs Addiction: Where the Line Actually Falls

A habit runs on autopilot. Addiction continues despite real harm and a genuine wish to stop. Here is the difference clinicians use, and what it means for you.

Do You Have to Hit Rock Bottom to Recover?

The idea that people only change once they have lost everything is folklore, not evidence. What actually predicts recovery, and why waiting makes it harder.

Latest research and news