Is Addiction a Disease?

The short answer

It can look like one, but it’s more of a syndrome. And you will get more out of the messy answer than the tidy one.

Addiction involves real, measurable changes in the brain's reward, stress, and self-control systems, which is what the disease model is pointing at, and it is why addiction belongs in a clinic or health setting. But "disease," as most people use the word in everyday speech, smuggles in things the evidence does not support. Permanence. A specific brain signature a scan could find. A clear biological determinant. Careful defenders of the model struggle to defend those implications too. Meanwhile, for the four substances where we have good national data, most people who become dependent eventually stop meeting the clinical criteria, though it often takes many years. My own position, after twenty years of studying this and a few years I would rather not repeat, is that the brain changes are real and are best understood as unusually deep learning. Whether you call that a disease is a decision about how to treat people. There is no scan for addiction, and a syndrome describes a collection of symptoms that show up together, but may have different causes - that’s a better fit.

What the evidence actually says

The disease model is usually argued against in a form nobody actually holds. Nora Volkow, George Koob, and A. Thomas McLellan set out the modern version in a 2016 review in the New England Journal of Medicine (DOI: 10.1056/NEJMra1511480; PMID: 26816013). Repeated heavy use produces lasting adaptations in the circuits handling reward, stress, and executive control. Those adaptations are why intention alone so often fails, and why relapse is so common. That is a fact about circuitry, not about character. When Volkow and Koob defended the model in Lancet Psychiatry in 2015 they pointed to the neuroplasticity evidence, to medications that work, and to insurance parity legislation as things the model produced (DOI: 10.1016/S2215-0366(15)00236-9; PMID: 26249284). That is a stronger case than the one usually attributed to them, and I want it on the table first.

Their point about medication is the one most readers should act on. For alcohol, Melissa McPheeters, Daniel Jonas, and colleagues published the current evidence base in JAMA in 2023: 118 clinical trials, 20,976 participants (DOI: 10.1001/jama.2023.19761; PMID: 37934220). The number needed to treat to prevent one person returning to any drinking was 11 for acamprosate (95% CI 1 to 32) and 18 for oral naltrexone at 50 mg/day (95% CI 4 to 32); for preventing a return to heavy drinking, naltrexone's number needed to treat was 11 (95% CI 5 to 41). They conclude that oral naltrexone at 50 mg/day and acamprosate are first-line pharmacotherapies alongside psychosocial treatment. Harms are real and worth knowing before you ask: more diarrhea with acamprosate, more nausea and vomiting with naltrexone. For opioids the stakes are higher. Thomas Santo and colleagues pooled 36 cohort studies covering 749,634 participants in JAMA Psychiatry in 2021 and found all-cause mortality during opioid agonist treatment, methadone or buprenorphine, at less than half the rate seen during time off treatment (RR 0.47) (DOI: 10.1001/jamapsychiatry.2021.0976; PMID: 34076676). In the four weeks after leaving treatment, mortality ran six times higher than during the rest of the time off it (RR 6.01). Two things belong beside that figure. It comes from observational cohorts, where treatment was not randomly assigned, so read it as a strong association. And the first four weeks on methadone carry roughly double the mortality of the rest of treatment (RR 2.01), which is exactly why induction is supervised and why dose decisions belong to a prescriber.

There is no diagnostic test for addiction. Neither is there one for a lot of diseases nobody argues about. Deborah Hasin and the DSM-5 substance-disorders work group laid out the current framework in a 2013 paper in The American Journal of Psychiatry, and their central recommendation, merging the old abuse and dependence categories into one substance use disorder, rested on consistent findings across more than 200,000 participants (DOI: 10.1176/appi.ajp.2013.12060782; PMID: 23903334). Diagnosis runs across eleven criteria. Two of them, tolerance and withdrawal, describe physiological states, though in practice a clinician establishes them by history. Craving is subjective. The rest are behavioral. Severity is graded by how many criteria you meet: two or three mild, four or five moderate, six or more severe, thresholds that were a committee's pragmatic judgment. So there is no blood value and no biopsy. Migraine, major depression, fibromyalgia, and irritable bowel syndrome do not have one either, and nobody stands outside a neurology clinic arguing that migraine is a lifestyle choice. The missing biomarker is interesting. It does not settle this question in either direction.

The empirical critique is sharper than most people expect, and it comes from inside the field. Wayne Hall, Adrian Carter, and Cynthia Forlini published an assessment in Lancet Psychiatry in 2015 that took the model on its own terms and asked whether it had delivered (DOI: 10.1016/S2215-0366(14)00126-6; PMID: 26359616). Their conclusions: the animal and neuroimaging evidence does not support the model to the extent its advocates suggest; it has not helped to deliver more effective treatments; and its effect on drug policy has been modest. They add a criticism that lands harder on me than the others. By concentrating on disordered neurobiology in a minority of severely addicted people, the model draws attention away from population-level measures that reduce heavy drinking and smoking cheaply and at scale. Volkow and Koob replied in the same journal that same year, arguing these claims are not supported by the evidence and pointing to the medications and the 2008 mental-health parity act (DOI: 10.1016/S2215-0366(15)00236-9). Read both. That exchange is the actual state of the debate, and neither side embarrasses itself.

Then there is the finding both camps have to account for. Catalina Lopez-Quintero and colleagues used the 2001–2002 NESARC national survey, 43,093 participants, to estimate lifetime cumulative probability of remission from dependence (DOI: 10.1111/j.1360-0443.2010.03194.x; PMID: 21077975). The estimates: 83.7% for nicotine, 90.6% for alcohol, 97.2% for cannabis on a subsample of 530 people, and 99.2% for cocaine on a subsample of 408. The caveats here are not decorative. These are lifetime cumulative-probability projections from a single survey wave, not observed proportions, and the authors note that the survey missed people who experienced fatal or severe consequences, which overestimates remission. "Remission" means no longer meeting DSM-IV dependence criteria, a different instrument from the DSM-5 severity bands above, and it is not the same as abstinence, recovery, or a good life. It is often very slow. Half of cases remitted roughly 26 years after onset for nicotine, 14 for alcohol, 6 for cannabis, 5 for cocaine. Remission was less likely for men across all four substances, less likely for Black respondents for nicotine and cocaine specifically, and less likely for people with personality disorders or more than one substance problem; lifetime totals varied by racial and ethnic group, so the pooled number predicts nothing about any group. NESARC surveyed adults, so it reaches no adolescents. And it contains no opioid stratum at all, which matters in a fentanyl era: there is no comparable national remission estimate for opioid dependence, so do not read these figures as one. Take all of that off the top and there is still a lot left. "Chronic and progressive with no exit" does not describe the population.

One more thing the data show that this debate rarely mentions. In that same survey, around 80% of people with nicotine or alcohol dependence had another lifetime psychiatric diagnosis, and nearly everyone with cannabis or cocaine dependence did. Much of that total is other substance disorders, and the figure for mood, anxiety, and personality disorders specifically is smaller, roughly a third for nicotine and alcohol. Either way, the practical point holds. No argument about labels is going to catch your anxiety disorder. An assessment will.

If you are physically dependent on alcohol, benzodiazepines, or barbiturates, do not stop abruptly. That withdrawal can kill you through seizures or delirium tremens. GHB is in the same danger class. Gabapentinoid and other sedative withdrawal can cause serious problems including seizures, and should also never be stopped abruptly on your own, though the evidence that it is commonly fatal is weaker. Signs you may be physically dependent: shakes in the morning, sweats, or needing a drink or a dose to stop symptoms starting. Call a clinician or urgent care today, not this week. Go to an emergency department, or call 911, for shaking you cannot control, heavy sweating, repeated vomiting, racing heart, fever, confusion, seeing or hearing things that are not there, or any seizure. "Do not stop cold" is not permission to keep drinking at your current level; a clinician can taper you safely, and that is the actual answer. On opioids: after time away, your tolerance drops, and the illicit supply in the US is now largely fentanyl, often with adulterants such as xylazine, so an old dose can be lethal. Never use alone. Keep naloxone on hand, be aware that fentanyl often needs more than one dose of it, and call 911 even after naloxone works. Naltrexone is an opioid blocker: it cannot be combined with methadone or buprenorphine and must not be started while opioids are in your system, so tell any prescriber if you are on opioid agonist treatment. Acamprosate needs a kidney-function check. If you are or might be pregnant, do not stop opioids, alcohol, or sedatives on your own and do not start or stop any of these medications without obstetric and addiction-medicine input; agonist treatment is the standard of care in pregnancy. The SAMHSA National Helpline is free and confidential, 24/7, at 1-800-662-4357, treatment options are at findtreatment.gov, and the 988 Suicide & Crisis Lifeline is there if you are in crisis. These are US resources.

Where the experts disagree

Everyone here agrees the brain is involved. The skeptics dispute how strong, how specific, and how clinically meaningful the imaging and animal findings are, which is a fight about the evidence itself.

The disease camp (Volkow, Koob, Heilig, NIDA). Neuroadaptations in reward, stress, and control circuitry undermine self-regulation (DOI: 10.1056/NEJMra1511480). Their case, in their own telling, runs neuroplasticity, then medications, then insurance parity (DOI: 10.1016/S2215-0366(15)00236-9). They also argue the framing itself reduces stigma and gives people hope of recovery. That last claim is contested, and I take it up below.

The revised defense (Heilig and colleagues, 2021). Worth reading because it concedes ground. Markus Heilig, James MacKillop, Diana Martinez, Jürgen Rehm, Lorenzo Leggio, and Louk Vanderschuren name the four standing criticisms directly: that the view is deterministic, that it fails to account for how much people differ in remission and recovery, that it over-emphasizes compulsion, and that no specific neural signature of addiction has been identified (DOI: 10.1038/s41386-020-00950-y; PMID: 33619327). They write that some of these criticisms have merit while holding the neurobiological premise sound. They do not say which ones, and I will not put words in their mouths. Two of their arguments are worth your time. Apparently spontaneous remission does not refute a brain-based account, because the brain is also the organ doing the changing, and they call for intensified neuroscientific study of recovery. And, as a claim about consequences, they argue that denying the brain disease view reduces access to healthcare. Their stated conclusion is a call for consilience: multidisciplinary research integrating neuroscientific, behavioral, clinical, and sociocultural perspectives. I should say plainly that this is close to the integrative position I land on myself, so I am not claiming much novelty here.

The empirical skeptics (Hall, Carter, Forlini). They do not dispute the neuroscience. They audit what it delivered. The model promised better treatments and better policy, and on their reading it delivered less than advertised while aiming attention at the most severe minority (DOI: 10.1016/S2215-0366(14)00126-6). Volkow and Koob's reply above is the other half of that exchange.

The learning camp (Marc Lewis). Lewis is a neuroscientist who was himself addicted, and his 2018 New England Journal of Medicine piece argues the same brain changes read better as deep learning than as disease (DOI: 10.1056/NEJMra1602872; PMID: 30332573). The brain doing what it evolved to do, wiring hard around something that reliably delivered relief. My phrasing, not his: a groove worn unusually deep, not an organ gone wrong. Grooves get re-cut. I have watched it happen, more than once, in people who had been told it was permanent. Note that this piece is a perspective review with no indexed abstract, so there is no summary to check it against.

The choice camp (Gene Heyman). Meet him in his actual form, which is not "addicts just choose to use." Heyman's 2013 paper in Frontiers in Psychiatry argues that the recovery data fit quantitative choice principles, the matching law, melioration, and hyperbolic discounting, better than they fit compulsion (DOI: 10.3389/fpsyt.2013.00031; PMID: 23653607). Most people meeting criteria for addiction to illegal drugs quit by about age 30, usually without professional help, and what correlates with quitting is legal trouble, money pressure, and wanting their family's respect. He goes further than most colleagues would, stating that the brain disease model is not supported by research or logic. Two honest notes belong here. His remission evidence comes from the same class of retrospective household surveys I just caveated, so the survivorship problem applies to him too. And Heilig and colleagues answer his mechanism directly, arguing that seemingly compulsive behavior can co-exist with sensitivity to alternative reinforcement, which means "it responds to incentives" does not by itself pick a winner (DOI: 10.1038/s41386-020-00950-y).

The pharmacology-and-policy critique (Carl Hart). People who have only heard Hart on a podcast are usually surprised that his most-cited contribution here is a peer-reviewed critical review. In a 2012 Neuropsychopharmacology paper he and his colleagues reviewed the cognitive literature on recreational methamphetamine users and concluded that statistically significant differences appeared on a minority of measures, that performance overwhelmingly fell within the normal range against normative data, and that there is a propensity to interpret any cognitive or brain difference as a clinically significant abnormality (DOI: 10.1038/npp.2011.276; PMID: 22089317). That challenges how such findings get read. It has also been answered: Stéphane Potvin and colleagues meta-analysed 44 studies covering 1,592 people with methamphetamine use disorder against 1,820 controls in 2018 and found moderate impairment across most cognitive domains, in the same range as alcohol and cocaine use disorder, while noting a publication bias (DOI: 10.1016/j.addbeh.2018.01.021; PMID: 29407687). Both are on the table. Hart also co-authored a 2021 American Journal of Bioethics paper arguing that drug criminalization was rooted in explicit racism, calling for decriminalization of all recreational drugs and eventually legal regulation (DOI: 10.1080/15265161.2020.1861364; PMID: 33413050). State his position fully or not at all. He is not asking for gentler language.

My read: they share most of the facts and fight over weight. The brain changes. The change is a form of learning. The behavior stays responsive to circumstance. The label does political work in the world, and it deserves auditing. If your account of addiction has to deny one of those four, it is the wrong account.

My take: this is a question about the map

I stopped arguing about this the day I noticed what the argument was actually about. "Is addiction a disease?" is a question about a word, and about what we want that word to do.

A word earns its keep by what it does. Any honest account has to explain three things at once, and most popular accounts pick one and drop the others.

  1. Why it is so hard to stop. People who love their kids and want to keep their jobs cannot simply decide their way out. The neuroadaptation literature explains this well.

  2. Why most people do stop. This is a test of the folk version of chronic disease, the one that circulates in meetings and family arguments rather than in journals. Remission is the majority long-run outcome for the substances we have good data on, so the folk version predicts the wrong thing about most people who have it. The careful defenders already agree, which is worth knowing when someone quotes "chronic and relapsing" at you as if it were a life sentence.

  3. Why circumstances move it. Money pressure, a new relationship, a move, a court date, a baby.

The framing that holds all three is the one my work is built on. Addiction is a learned solution to a problem, usually an emotional one, that worked so reliably your brain kept reaching for it after the costs overtook the benefits. Learning is biological, so the brain changes are real. It is also revisable, which is why people recover, and it bends to context, which is why a move or a new job or a baby can change everything.

My SPARO chain is how I trace it with someone in a room. SPARO is Stimulus, Perception, Activation, Response, Outcome. A Stimulus lands: a text, a Sunday afternoon, the sound of your own front door. Your Perception reads it as threat or lack. That produces an Activation, the felt charge, the tightness in your chest, the hollow behind your ribs. The Response is whatever most reliably switches that charge off, the drink, the pill, the bet, the scroll. The Outcome is relief, and relief teaches the chain to run faster next time. Do it enough and the Response fires before you have consciously decided anything. That is what "loss of control" feels like from the inside. The scan is gesturing at the same thing from a long way off.

Which gives you something practical. Use the disease frame when it does the work it is good at, getting you covered and getting you treated. It also takes some weight off your chest. Set it down when it does the harm it is prone to, convincing you that you are a permanently broken machine waiting for the next breakdown. No laboratory handed this word down. It is a tool, and you are allowed to pick it up and put it down.

One thing the tool does not decide for you. Medication is compatible with every framing on this page, mine included. If you have opioid or alcohol dependence, the medications are among the best-evidenced things in the entire field, and the mortality data on opioid agonist treatment are not a close call (Santo et al., 2021).

I wrote The Abstinence Myth partly because I watched the word "disease" comfort one person in a room and quietly sentence the person next to him. Same word. Same night. Nothing in the research changed between those two men. Only the sentence they heard did.

What to actually do

✅ Get the safety part settled first, before any experiment. If you drink heavily every day, or use benzodiazepines, barbiturates, or GHB regularly or on prescription, withdrawal is a medical event and nobody gets points for toughing it out. Call SAMHSA at 1-800-662-4357 or talk to a clinician before you change anything. If you are or might be pregnant, that conversation comes before every other item on this list.

✅ Ask about medication, whatever you decide about the word. For alcohol dependence, oral naltrexone at 50 mg/day and acamprosate are first-line and underprescribed (McPheeters et al., 2023). For opioid dependence, methadone and buprenorphine are associated with roughly half the all-cause mortality of time off treatment, and mortality is about six times higher in the four weeks after leaving it (Santo et al., 2021). Tell any prescriber if you are on opioid agonist treatment, because naltrexone cannot be combined with it. And if you are already on medication, do not come off it because a philosophical argument persuaded you.

✅ Judge the label by what it does for you. If "disease" gets you into treatment and gets you to stop hating yourself, use it and do not apologize. If it makes you feel doomed and passive, put it down. Both responses are common, and neither is scientifically illiterate.

✅ Ask your clinician what their model predicts. Ask it warmly. "What does your framing say about my prognosis, and what would change it?" is a good conversation. One trap in the other direction: for opioid or alcohol dependence, "chronic, needs long-term treatment" is often the clinically correct answer, and that is usually why they want you to stay on the medication.

✅ Get properly assessed for what else is going on. Most people with dependence in the national data carried at least one other lifetime psychiatric diagnosis. A label argument cannot find your depression.

✅ Whichever word you use, ask the SPARO question. For one week, notice the feeling in the minute before the urge. Most people have never looked there. That feeling is the problem, and the behavior is your current answer to it.

✅ Refuse both extremes when you talk to family. "It's a disease so I can't help it" and "it's a choice so just stop" both end conversations. The accurate middle is harder and more useful. This is a deeply learned response, genuinely hard to interrupt, and it genuinely does get interrupted, and I need support that treats me as capable.

✅ If you are worried about someone else, especially a teenager, do not use this page's numbers. The remission data come from a survey of adults and say nothing about adolescents, who need assessing now. Waiting and watching is what costs you the year. Family-focused approaches have their own evidence base. Keep naloxone accessible if opioids are anywhere in the picture, know that fentanyl often needs more than one dose, and use 988 or 911 for the acute moments.

FAQ

Is addiction a disease or a choice? A false binary, and both camps have a point. The brain genuinely changes (Volkow et al., 2016), those changes look a lot like deep learning (Lewis, 2018), and the behavior stays responsive to incentives and circumstance (Heyman, 2013). The useful question is what the behavior is doing for the person.

Is there a brain scan or blood test that can diagnose addiction? No. Diagnosis runs across eleven clinical criteria: two physiological, one subjective, the rest behavioral (Hasin et al., 2013). "No specific neural signature has been identified" is one of the standing criticisms the model's own 2021 defenders engage with (Heilig et al., 2021). Hart's review argues such differences are routinely over-read as clinical abnormality (Hart et al., 2012); a later meta-analysis found moderate impairment across most cognitive domains in methamphetamine use disorder (Potvin et al., 2018). Imaging shows differences between groups of hundreds of people. It cannot tell a clinician anything about the person in the chair.

If addiction is a disease, why do most people recover without treatment? That is the objection, and the disease camp has an answer worth hearing: the brain is also the organ that changes, so remission is itself something the brain does (Heilig et al., 2021). Whether that reads as persuasive or circular is roughly where you land. On the data: lifetime cumulative remission from DSM-IV dependence is estimated at 83.7% for nicotine, 90.6% for alcohol, 97.2% for cannabis, and 99.2% for cocaine, with medians of about 26, 14, 6, and 5 years, small subsamples for cannabis and cocaine, no opioid estimate, and household-survey exclusions the authors say overestimate remission (Lopez-Quintero et al., 2011). They urge caution given how irregular the course of addiction is, and so do I.

Does calling addiction a disease reduce stigma? It was meant to, and whether it did on net is genuinely disputed. Volkow and Koob argue the framing diminishes stigma and gives hope of recovery (Volkow & Koob, 2015); Hall and colleagues explicitly set out to assess claims about the model's social benefits and found them weaker than advertised (Hall et al., 2015). The wider research on biological explanations of mental illness points both ways: blame tends to fall, while pessimism about recovery can rise. Medical framing lowers one burden and can raise another, the sense of being permanently defective.

Is alcoholism a disease? Same answer as the general case, with one addition that is not philosophical. Alcohol is one of the few drugs whose withdrawal can kill you, and effective medications exist for it (McPheeters et al., 2023). Whatever you conclude about the word, physical dependence on alcohol is a medical situation needing a clinician before you make changes.

Does it matter what I believe about this? More than you would think. What you believe shapes whether you ask for help, how you read a lapse, and whether you treat yourself as a project or a patient. It does not change the mechanism, and it should never be the reason you decline an effective medication.

Related in the Behavior Change Atlas

References

Citations verified against PubMed; sources last checked 2026-07-30.

  1. Volkow ND, Koob GF, McLellan AT. Neurobiologic advances from the brain disease model of addiction. N Engl J Med. 2016;374(4):363–371. DOI: 10.1056/NEJMra1511480 PMID: 26816013

  2. Volkow ND, Koob G. Brain disease model of addiction: why is it so controversial? Lancet Psychiatry. 2015;2(8):677–679. DOI: 10.1016/S2215-0366(15)00236-9 PMID: 26249284

  3. Heilig M, MacKillop J, Martinez D, Rehm J, Leggio L, Vanderschuren LJMJ. Addiction as a brain disease revised: why it still matters, and the need for consilience. Neuropsychopharmacology. 2021;46(10):1715–1723. DOI: 10.1038/s41386-020-00950-y PMID: 33619327

  4. Hall W, Carter A, Forlini C. The brain disease model of addiction: is it supported by the evidence and has it delivered on its promises? Lancet Psychiatry. 2015;2(1):105–110. DOI: 10.1016/S2215-0366(14)00126-6 PMID: 26359616

  5. Lewis M. Brain change in addiction as learning, not disease. N Engl J Med. 2018;379(16):1551–1560. DOI: 10.1056/NEJMra1602872 PMID: 30332573

  6. Heyman GM. Addiction and choice: theory and new data. Front Psychiatry. 2013;4:31. DOI: 10.3389/fpsyt.2013.00031 PMID: 23653607

  7. Hart CL, Marvin CB, Silver R, Smith EE. Is cognitive functioning impaired in methamphetamine users? A critical review. Neuropsychopharmacology. 2012;37(3):586–608. DOI: 10.1038/npp.2011.276 PMID: 22089317

  8. Potvin S, Pelletier J, Grot S, Hébert C, Barr AM, Lecomte T. Cognitive deficits in individuals with methamphetamine use disorder: a meta-analysis. Addict Behav. 2018;80:154–160. DOI: 10.1016/j.addbeh.2018.01.021 PMID: 29407687

  9. Earp BD, Lewis J, Hart CL, et al. Racial justice requires ending the war on drugs. Am J Bioeth. 2021;21(4):4–19. DOI: 10.1080/15265161.2020.1861364 PMID: 33413050

  10. Hasin DS, O'Brien CP, Auriacombe M, et al. DSM-5 criteria for substance use disorders: recommendations and rationale. Am J Psychiatry. 2013;170(8):834–851. DOI: 10.1176/appi.ajp.2013.12060782 PMID: 23903334

  11. Lopez-Quintero C, Hasin DS, de Los Cobos JP, et al. Probability and predictors of remission from life-time nicotine, alcohol, cannabis or cocaine dependence: results from the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC). Addiction. 2011;106(3):657–669. DOI: 10.1111/j.1360-0443.2010.03194.x PMID: 21077975

  12. McPheeters M, O'Connor EA, Riley S, et al. Pharmacotherapy for alcohol use disorder: a systematic review and meta-analysis. JAMA. 2023;330(17):1653–1665. DOI: 10.1001/jama.2023.19761 PMID: 37934220

  13. Santo T, Clark B, Hickman M, et al. Association of opioid agonist treatment with all-cause mortality and specific causes of death among people with opioid dependence: a systematic review and meta-analysis. JAMA Psychiatry. 2021;78(9):979–993. DOI: 10.1001/jamapsychiatry.2021.0976 PMID: 34076676

Limitations stated honestly: This page describes a live scientific and philosophical disagreement and takes an integrative position rather than declaring a winner; reasonable experts land in different places. Several key sources are argued reviews, not experiments: Volkow et al. (2016), Volkow & Koob (2015), and Heilig et al. (2021) argue for the brain disease model; Hall et al. (2015) argues against its evidentiary reach and its record; Lewis (2018) and Heyman (2013) argue for learning and choice framings. Three of those — Volkow et al. (2016), Volkow & Koob (2015), and Lewis (2018) — have no indexed PubMed abstract, so their existence was verified there and their content read in full text. No trial settles this question, because none can. The remission figures (Lopez-Quintero et al., 2011) are lifetime cumulative-probability projections from a single 2001–2002 survey wave using retrospectively recalled ages of onset and remission; they concern DSM-IV dependence rather than DSM-5 severity bands, exclude adolescents and people who were incarcerated, homeless, institutionalized, or deceased, rest on subsamples of 530 and 408 for cannabis and cocaine, vary substantially by racial and ethnic group, and contain no opioid estimate. The same survivorship and recall limits apply to Heyman's remission evidence. Whether medical framing reduces stigma on net is disputed rather than settled, and the biological-attribution literature points both ways. The DSM-5 criteria are a clinical consensus instrument, not a biological test, and the severity thresholds were a pragmatic committee judgment; DSM-5-TR (2022) is the current edition. The medication meta-analyses (McPheeters et al., 2023; Santo et al., 2021) report group-level benefit and cannot predict an individual response, and the Santo estimates come from observational cohorts where treatment was not randomly assigned. The Hart (2012) and Potvin (2018) papers concern methamphetamine cognition specifically and should not be generalised to the whole neuroimaging literature; Potvin also reports a publication bias. Finally, "addiction as a learned solution to an emotional problem," the three-test frame, and SPARO are my own clinical and personal framing built on this evidence rather than the finding of any single study. Nothing here is individual medical advice, and heavy use or likely physical dependence warrants professional guidance before you change anything, including any change to prescribed medication.

Written by Dr. Adi Jaffe, PhD (UCLA), author of The Abstinence Myth and Unhooked. I am a psychologist, not a physician; nothing here replaces care from a prescribing clinician. Disclosure: I am the founder of IGNTD, which provides paid behavior-change and addiction programs; no IGNTD product is recommended on this page.