What the West Got Wrong About Drug Decriminalization: Lessons From Portugal and Oregon

This post was originally published on this site.

What the West Got Wrong About Drugs

Twenty-five years ago, Portugal did something that should have gotten its health minister laughed out of every G7 summit: it decided that a person caught with a small amount of heroin was a patient, not a criminal. This year marks the quarter-century anniversary of that decision, and the retrospective published via The Conversation and picked up by Marijuana Moment gives me a good excuse to ask the question I keep circling back to: what did the West actually get wrong, and is any of it fixable?

I’ll answer with the numbers first and the arguing second, because that’s the order that keeps me honest.

Before 2001, Portugal was in genuine crisis. An estimated 1 percent of the population had a heroin problem, and the country had the highest HIV incidence in the European Union among people who inject drugs. That is not a society managing drug use. That is a society losing to it.

In 1999, a commission of physicians, psychiatrists, and legal experts designed a new national strategy. Portugal decriminalized personal possession of all drugs in 2001, meaning you cannot be jailed for having a small quantity, though you can still be referred to a “dissuasion commission” that can connect you to treatment or issue minor sanctions. Manufacturing and trafficking stayed fully criminal. This is a distinction people flatten constantly, so I want to be precise: Portugal did not legalize heroin. It stopped prosecuting the sick for being sick.

The results, per the researchers who’ve spent three years teaching a course on the ground there:

New HIV diagnoses tied to injection drug use fell from 583 in 2005 to just 19 in 2024, a decline of roughly 97 percent.
Overdose deaths, which had been climbing sharply before the policy, fell and have stayed low. In 2023, Portugal recorded 105 fatal overdoses. Note that this is up slightly from 96 in 2022 and 81 in 2021, and I’m not going to pretend that trend line is flat. It isn’t. But it remains far below the crisis-era numbers and far below what the U.S. sees per capita.
More than 24,000 people received care through the public outpatient treatment system in 2023 alone.

Now the part advocates skip, and I won’t: Portugal’s success wasn’t the decriminalization law by itself. It was decriminalization plus a financial commitment to treatment infrastructure that survived recessions and budget cuts for a quarter-century. And even Portugal isn’t a fairy tale ending: Porto’s mayor has publicly complained about visible public drug use downtown, and urban dissuasion commissions report being underfunded relative to need. Success in the aggregate national data doesn’t mean zero friction on the ground.

If Portugal is the case for treating drugs as a health problem, Oregon is the control group that should make you nervous, because Oregon tried something similar and it detonated politically. Measure 110 decriminalized possession in 2020, funded by cannabis tax revenue, and Oregon recriminalized it in 2024, four years later.

Here’s where I have to slow down, because the causal story got mangled in the press and I don’t want to repeat that mistake. The politically dominant narrative was: decriminalization caused an overdose spike, therefore decriminalization failed. The peer-reviewed evidence says something considerably messier.

A 2023 JAMA Psychiatry study using synthetic control methods found no statistically significant association between Measure 110’s implementation and Oregon’s fatal overdose rate in year one. A 2024 JAMA Network Open study that explicitly modeled fentanyl’s arrival found the same thing: once you account for the fact that fentanyl was flooding Oregon’s drug supply at almost exactly the same moment the law took effect, the “M110 caused this” story loses its legs. A 2026 follow-up using CDC WONDER mortality data found Oregon’s death-rate inflection point was December 2019, fourteen months before the law even took effect, and that neighboring states with no decriminalization law saw nearly identical shifts on nearly identical timelines.

I’ll also give the other side its due, because pretending there’s no dissent isn’t honesty, it’s advocacy in a lab coat. At least one earlier analysis (Spencer, 2023) attributed a 23 percent increase in overdose deaths to Measure 110, and a newer synthetic-control paper argues the standard fentanyl adjustment may itself be absorbing part of a real policy effect. The academic literature genuinely disagrees on the size of the effect. What it does not support is the simple version everyone actually believed: that decriminalization alone drove the crisis.

So why did Oregon fail politically where Portugal succeeded? The Marijuana Moment retrospective is blunt about it: Oregon decriminalized without building Portugal’s connective tissue. Treatment funding was delayed. Law enforcement was never brought in as a partner and stayed openly hostile to the policy. New services bolted onto an already-fragmented behavioral health system instead of forming an integrated one. Highly visible public drug use fed a media and political backlash that outran the actual data. Oregon lawmakers repealed the law in 2024 before the state had even finished collecting reliable outcome data to judge it by.

British Columbia’s 2023 decriminalization pilot sits in between: Canada has universal healthcare and a mature harm-reduction infrastructure, but police weren’t required to refer people to services and treatment providers got no new funding. Possession offenses and seizures dropped sharply; broader health benefits didn’t materialize, and the province let the pilot expire in January 2026.

The pattern across all three is not “decriminalization works” or “decriminalization fails.” It’s that removing criminal penalties without simultaneously building the health system to catch people is decriminalization in name only, and it collapses under its own vacuum the first time public disorder becomes visible on a sidewalk.

If Portugal decriminalized possession, Switzerland actually prescribed the drug. Starting in 1994, Switzerland began giving pharmaceutical-grade heroin, under medical supervision, to people with severe, treatment-resistant opioid addiction, as one part of a broader “four pillars” strategy of harm reduction, treatment, prevention, and enforcement together.

The results, tracked over decades by multiple independent reviews including a 2011 Cochrane Collaboration review, a widely-cited academic case study, and program summaries spanning 1994 to 2025:

Overdose deaths dropped an estimated 50 to 64 percent, depending on the study window.
HIV infections among participants dropped roughly 65 to 84 percent.
No participant enrolled in the heroin-assisted treatment program has died of a heroin overdose since the program began, according to program summaries — the treatment is medically supervised, dosed, and administered with naloxone on hand.
Felony crime among participants dropped roughly 60 percent, with an 82 percent drop in participants dealing heroin to fund their own habit, per the same program data.
New heroin initiation actually declined. The theory that supervised, “boring” medical heroin makes the drug less attractive to young people, rather than more, is explicitly noted in a 2006 Lancet paper and reaffirmed by Swiss health officials as recently as 2025.

This is the fentanyl-era question you asked, and it deserves a straight answer: yes, Switzerland’s model is essentially “compete with the black market on purity and reliability,” and yes, the data says it works at the individual and public-health level. It is not a free-for-all. Enrollment requires documented treatment failure on other options, daily supervised dosing at a clinic, and wraparound psychiatric and social support. It costs money and infrastructure, same as Portugal.

I’d flag one live cautionary tale from next door: British Columbia’s “safer supply” program, which prescribes take-home hydromorphone rather than requiring supervised on-site dosing, has run into a real diversion problem. A 2025 clinical study found that of 50 patients tested, only 62 percent were taking their hydromorphone as prescribed; the rest were selling, trading, or using it in ways other than prescribed, though the same study found zero overdose deaths among those 50 patients over the study year. B.C. has since moved toward mandatory witnessed dosing, and enrollment dropped by more than half as a result. This is the actual, unglamorous lesson: unsupervised take-home dosing invites diversion; supervised on-site dosing, Switzerland’s actual model, does not have the same problem. The details of implementation are not a footnote. They’re the entire ballgame.

You raised something from your own time walking among unhoused populations in California, and the research backs the instinct, with an important correction to the emphasis. The 2023 UCSF California Statewide Study of homelessness, the largest such study since the mid-1990s, found that housing cost, not addiction or mental illness, was the dominant driver: participants had a median household income of just $960 a month before losing housing. But the behavioral health burden once people are unhoused is severe and real: two-thirds reported current mental health symptoms, and a large share reported regular substance use, with methamphetamine, not opioids, the most commonly reported drug. Crucially, one in five people who wanted substance use treatment said they couldn’t get it.

That’s the whole argument in one statistic. It’s not that people don’t want help. It’s that the help isn’t there, and criminalizing them in the meantime doesn’t manufacture capacity that doesn’t exist.

Not that punishment is always useless; trafficking enforcement remains part of every model here, Portugal’s included. What got it wrong was believing that criminal law could substitute for a health system that was never built, and that removing the criminal law without building that system would produce the same result as removing it after building the system. Portugal spent 25 years building the system. Oregon spent four years without one and called the experiment a failure. Switzerland went furthest, treated addiction as a chronic medical condition requiring supervised medical management, and got the strongest numbers of the three.

None of this is a slogan. It’s slow, funded, unglamorous infrastructure, sustained past the point where it’s politically convenient. That’s a much harder sell than “decriminalize” or “legalize,” which is probably why the West keeps reaching for the easy half of the policy instead of the whole thing.

UCSF, California Statewide Study of People Experiencing Homelessness

DID PORTUGAL WORK, READ ON…

DID PORTUGAL WORK FOR DRUG REHAB? READ THIS!

DID PORTUGAL WORK, READ ON…

Leave a Reply Cancel reply