Industries / Health / Healthcare / Behavioral health
Addiction treatment
Helping people stop or cut down harmful drug and alcohol use, and keeping them alive meanwhile, through mutual help, medicines, counseling and harm reduction, under laws that shape all of these.
7
Milestones achieved, from 1935 to 2024.
2
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1935
Alcoholics help each other
On 12 May 1935, in Akron, Ohio, Bill W., a New Yorker in town on business, met Dr. Bob S., a local surgeon who drank. The date of Dr. Bob's last drink, 10 June 1935, is celebrated as the founding date of Alcoholics Anonymous (AA), a free fellowship in which people with drinking problems help each other stay sober. A 2020 Cochrane review of 27 studies with 10,565 participants found high-certainty evidence that structured counseling that steers people into AA kept more people continuously abstinent at 12 months than other treatments such as cognitive behavioral therapy (risk ratio 1.21).
Credit
Bill W. and Dr. Bob S.; J. F. Kelly, K. Humphreys and M. Ferri, for the review
1965
Methadone as maintenance
Vincent Dole and Marie Nyswander reported in JAMA in 1965 a clinical trial of methadone, a long-acting opioid taken once a day, as a steady maintenance treatment for heroin addiction in place of withdrawal. US rules from 1972 and the Narcotic Addict Treatment Act of 1974 confined methadone for addiction to licensed opioid treatment programs. A 2026 policy history says this "closed" clinic system clustered clinics in cities and separated addiction care from the rest of medicine.
LessonWhere a medicine may be given shapes who can get it.
Credit
Vincent P. Dole and Marie Nyswander
2001
Portugal decriminalizes drug use
Portugal's Law 30/2000, in force from 1 July 2001, kept using or possessing drugs illegal but moved it from the criminal courts to administrative procedures with health and social measures. A 2025 paper summarizing later research says the policy did not change how many people use drugs, and was followed by less problematic use, fewer drug-related harms and more people staying in treatment. A 2017 study found drug prices did not fall afterward.
Credit
Portugal's government and parliament
2002
Buprenorphine in the doctor's office
A US law of 2000, DATA 2000, let doctors prescribe approved opioid-addiction medicines from their own offices, and in 2002 the FDA approved two forms of buprenorphine, Subutex and Suboxone. Doctors needed a special "X-waiver" from the Drug Enforcement Administration (DEA), with 8 hours of training and a cap of 30 patients, later raised to 100 and 275. In 2012 only 16% of psychiatrists held one, and nearly 30 million Americans lived in counties without buprenorphine. A law signed on 29 December 2022 removed the waiver, and from 31 December 2025 a DEA rule lets clinicians start buprenorphine by telemedicine, including audio-only calls, for an initial supply of up to six months.
LessonA license meant to make prescribing safe kept a medicine scarce for two decades.
Credit
US Congress; FDA; DEA and the Substance Abuse and Mental Health Services Administration (SAMHSA)
2017
An app cleared as treatment
On 14 September 2017 the FDA granted Pear Therapeutics' reSET, a phone app that teaches cognitive behavioral therapy lessons and rewards clean urine tests, as a prescription device for substance use disorders. In the trial behind it, run by the National Institute on Drug Abuse's Clinical Trials Network, 40.3% of patients who did not mainly use opioids were abstinent in weeks 9 to 12 with the app plus counseling, against 17.6% with counseling alone; patients who mainly used opioids did no better. Pear filed for bankruptcy on 7 April 2023, and a 2025 commentary says insurers covered only about half of its prescriptions. A 2026 trial of the opioid version, reSET-O, in 100 patients at a US clinic found no significant effect on staying in treatment.
LessonClearance by a regulator does not make insurers pay for a treatment.
accessdata.fda.govaccessdata.fda.govsec.goveuropepmc.orgeuropepmc.org
Credit
Pear Therapeutics; NIDA Clinical Trials Network
2023
Naloxone without a prescription
Naloxone reverses an opioid overdose. On 29 March 2023 the FDA approved Narcan, Emergent BioSolutions' 4 mg naloxone nasal spray, for sale without a prescription, the first naloxone product approved for that. The approval used a picture-based Drug Facts label that the FDA had designed and tested in 2019, and a simulated-use study showing that ordinary people could use the spray correctly without a health professional.
Credit
Emergent BioSolutions; US FDA
2024
Oregon brings back penalties
Oregon's Measure 110, passed by voters in 2020 and in force from 1 February 2021, removed criminal penalties for possessing drugs for personal use and paid for more services. A 2023 study in JAMA Psychiatry found no change in fatal overdose rates in its first year compared with similar states. In 2024 the legislature passed House Bill 4002, which made possession a crime again from 1 September 2024, as a "drug enforcement misdemeanor", and paid for programs that steer people to treatment instead of jail.
LessonA policy can be reversed before its effects have been measured.
Credit
Oregon voters; Oregon Legislature
Watch for
Ibogaine for opioid addiction
Ibogaine, a psychoactive substance from the West African iboga shrub, is illegal under US federal law, and its backers say it can interrupt opioid addiction. After the Texas Ibogaine Initiative passed in 2025, the Texas Health and Human Services Commission awarded $50 million to a consortium of state universities, co-led by UTHealth Houston and the University of Texas Medical Branch, for a two-year program of trials called IMPACT that includes opioid use disorder.
Achieved whena randomized trial publishes results showing that ibogaine helps people with opioid use disorder with acceptable safety.
Claimed by
the Texas Ibogaine Initiative and the State of Texas, award reported January 2026
Diabetes drugs against drinking
GLP-1 receptor agonists, the class of drugs used for diabetes and obesity, are being tested against alcohol use disorder. A 2026 review by S. Woo, H. R. Kranzler and colleagues found "convergent evidence from animal and human studies" that they reduce drinking, but only 4 published trials, and said more randomized trials are needed. A count of ClinicalTrials.gov on 1 September 2026 found 34 ongoing randomized trials of these drugs in psychiatry and addiction, 14 of them in alcohol use disorder.
Achieved whena phase 3 randomized trial shows less heavy drinking than placebo.
Claimed by
Woo, Kranzler and colleagues, 2026, and the trials' sponsors
Open
01
Time the early warning
For each new adulterant that has spread through North American street drugs since 2020, such as xylazine, medetomidine and nitazene opioids, a city-by-city table of the date a drug-checking service first found it, the date it first appeared in public death or poisoning records, and the gap between the two.
The briefCredit · unclaimed02
One table per consumption site
An open dataset with one row per supervised consumption site per year, worldwide: visits, unique clients, overdoses attended, deaths on site and referrals, in one format, with every figure linked to its source and notes where definitions differ.
The briefCredit · unclaimed
Open 01 · the brief
Time the early warning
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
For each new adulterant that has spread through North American street drugs since 2020, such as xylazine, medetomidine and nitazene opioids, a city-by-city table of the date a drug-checking service first found it, the date it first appeared in public death or poisoning records, and the gap between the two.
Why it is still open
Drug-checking services report first detections one city at a time. Toronto's service first found xylazine in September 2020, and a 2026 preprint dates medetomidine's first detection in a Southern California program to May 2025, rising to about 25% of fentanyl samples by August 2026. A search found no study that measures, across cities and substances, how far drug checking runs ahead of death records.
Built with what exists
public drug-checking dashboards and reports from cities such as Toronto, Vancouver and Philadelphia; medical examiner and coroner reports; a web agent such as Browser Use to gather them (see the Note "Deleting 80% to start over"); Python. Nothing is handled but published data.
First step
Build the rows for xylazine in Toronto and Philadelphia from published reports and dashboards.
Done looks like
At least 4 substances in at least 6 cities, each first-detection and first-death date with a source link; the median lead time with its range; the table and method published.
Open 02 · the brief
One table per consumption site
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
An open dataset with one row per supervised consumption site per year, worldwide: visits, unique clients, overdoses attended, deaths on site and referrals, in one format, with every figure linked to its source and notes where definitions differ.
Why it is still open
Canada's federal dashboard gives standardized per-site figures, but only for Canada and only since March 2020, when standard reporting began. Elsewhere the figures sit in separate annual reports and evaluations. A September 2025 review by the Public Health Agency of Canada did not pool studies of population-level overdose deaths "because study design, exposure and outcome measures varied significantly". A search found no international per-site dataset.
Built with what exists
Health Canada's downloadable per-site CSV; national and site reports from Europe, Australia and the US; a language model that reads reports in several languages, such as the open-weight Mistral Large 4 (see the Note "A trillion open parameters"); Python and a public repository.
First step
Download Health Canada's per-site CSV and add the latest yearly figures for Sydney's Medically Supervised Injecting Centre and OnPoint NYC by hand to test the format.
Done looks like
At least 100 sites in at least 8 countries with at least three years each; 10% of figures rechecked by a second person with at least 98% agreement; a published codebook that maps each country's definitions.