# Lesson 12: Alternatives to Prohibition and an Integration of the Course

*Companion-podcast transcript • Sarah & Kiffer*  
*Office Hours episode to listen to after working through the lesson*

---

**Sarah:** Welcome back to Office Hours. I'm Sarah.

**Kiffer:** And I'm Kiffer. Lesson twelve compares the alternatives to prohibition and pulls the course together. We'll be discussing overdose deaths, and support lines are listed on the lesson page. One of those alternatives was tried in British Columbia, and on the fourteenth of January, 2026, the province announced that it would not seek to renew its decriminalization exemption.

**Sarah:** And two studies published in 2025 found no significant change in deaths. So did it work?

**Kiffer:** That depends on what you expected it to do. If the goal was fewer people facing police and criminal records for a small amount of a drug, the police data show a large change: while the exemption applied in full, police-recorded possession offences in the province fell by about two thirds. If the goal was fewer deaths, those two studies didn't find one, and a change in the penalty for possession leaves the content of the drugs exactly as it was.

**Sarah:** Question one. If decriminalization leaves the drug supply untouched, what is it supposed to change? I think a lot of people hear decriminalization and legalization as the same thing.

**Kiffer:** They're easy to blur, and the difference shapes how to read every result in this area. Decriminalization changes what happens to the person holding a small amount. Production and sale stay crimes, so that person still buys from the unregulated market. What it can be expected to change is contact with police, criminal records and their effects on jobs and housing, and perhaps people's willingness to seek care. British Columbia named reducing the stigma around getting care as one of its purposes.

**Sarah:** And on police contact, the province's own data are clear. Police-recorded possession offences fell from an average of five hundred nine a month to one hundred sixty-five. Then, after public spaces were excluded in May 2024, they climbed back to four hundred twenty-three.

**Kiffer:** Right, and the details of a design decide how much of that change happens. Take the threshold, the amount below which possession counts as personal use. British Columbia asked Health Canada for a combined limit of 4.5 grams, and in May 2022 Health Canada approved 2.5 grams.

**Sarah:** And that lower number was contested at the time. People who use drugs, advocates and some public health experts told reporters that 2.5 grams was too low for people who use often or buy larger amounts at once. Police had opposed the higher, per-drug thresholds that advocates wanted, and British Columbia's association of police chiefs had recommended a limit of one gram in total.

**Kiffer:** So the threshold reflected competing views among interest holders, and it set the size of the group the policy protected. A person carrying more than 2.5 grams in total was still exposed to arrest and charges.

**Sarah:** Oregon built a services pathway into its design. Under Measure 110, possession of a small amount became a violation with a fine of up to one hundred dollars, dismissed if the person completed a health screening through a hotline. Oregon's courts recorded ten thousand twenty-eight cases with these violations for offences from February 2021 to August 2024, and in eighty-five of them a verification of the screening was filed. Here's a question for everyone. Roughly what share of the cases is that? Pause the audio if you'd like more time. Take a few seconds.

*(Pause)*

**Kiffer:** Eighty-five divided by ten thousand twenty-eight is about 0.85 percent, so fewer than one case in a hundred. For comparison, seven hundred four cases, about seven percent, ended with all the charges dismissed, and the eighty-five are among those. Seven thousand nine hundred twenty-five, about seventy-nine percent, ended in a conviction after the person failed to appear in court.

**Sarah:** So the hotline route almost never ran the way it was designed to, at least through the courts.

**Kiffer:** That's the careful way to put it. The court data don't capture treatment the measure paid for outside that pathway, and Joshi and colleagues noted that the new treatment money went out slowly, mostly after the first year. Portugal took a different route. Its reform came with a large expansion of treatment, from about twenty-three thousand seven hundred people in treatment in 1998 to about thirty-eight thousand five hundred in 2008, and people found with small amounts are referred to regional commissions.

**Sarah:** So decriminalization is a family of designs. The threshold, the sanction and the services attached all decide what it can change.

**Kiffer:** And none of them changes what's in the drug. Only legal regulation, or a prescribed supply of the kind HSCI 402 examines, does that. When someone predicts that decriminalization alone will bring overdose deaths down quickly, it's worth asking what they think the mechanism is.

**Sarah:** Question two. Why do careful studies of the same policy reach different conclusions? Oregon has several analyses of Measure 110 and overdose deaths, and they don't line up.

**Kiffer:** The main reason is timing. Fentanyl spread rapidly into Oregon's supply in the first half of 2021, the same months the measure took effect, so every study has to pull apart two changes that started together.

**Sarah:** And the studies handled that differently. Spencer, using a synthetic control, estimated one hundred eighty-two additional deaths in 2021, about twenty-three percent more than predicted. Joshi and colleagues found no significant association in the first year. Zoorob and colleagues found an association that disappeared once they adjusted for fentanyl's spread. And then there's the 2026 paper.

**Kiffer:** Yes, a working paper by Hall, Hansen and Matsuzawa, released in July 2026, which, as of October 2026, hasn't been peer reviewed. It estimated about one thousand one hundred eighty-six excess deaths in Oregon from 2021 through 2023, meaning deaths above what its comparison predicted. It also questioned the adjustment itself. One common measure of fentanyl's spread is its share of drugs seized by police and sent for testing.

**Sarah:** So if police seized fewer of the other drugs once enforcement eased, fentanyl's share would go up partly because of the policy.

**Kiffer:** That's their argument. Adjusting for that measure could then hide part of the policy's effect, and they reported that other measures of fentanyl's spread still left a positive effect on deaths. It's a fair question to ask of any adjustment: could the policy itself have changed the thing you're adjusting for? Until that paper is reviewed and others test it, I'd describe the Oregon evidence as conflicting.

**Sarah:** British Columbia has a different pattern. The two 2025 studies agree that deaths didn't change significantly, and they disagree about hospitalizations.

**Kiffer:** Let's put numbers on one of them. Gaudreault and colleagues reported average opioid hospitalization rates per one hundred thousand people per quarter, a quarter being three months. In British Columbia, the rate was 9.05 in 2021 and 2022, and 10.34 in 2023, the exemption's first year. In the rest of Canada, it went from 2.95 to 3.03.

**Sarah:** Let me run it. British Columbia went up by 1.29 and the rest of Canada by 0.08, so the difference-in-differences is 1.21 per one hundred thousand per quarter. On the relative scale, 10.34 divided by 9.05 is about 1.14, and 3.03 divided by 2.95 is about 1.03. Dividing one by the other gives about 1.11. So on that quick calculation, British Columbia came out roughly eleven percent higher than the comparison would predict.

**Kiffer:** Your arithmetic is right for two averages, and it's a useful rough check. The study's own estimate came from a model fitted to the rate in each of the twelve quarters, and that model put the rise at about thirteen percent in British Columbia and about nine percent in the rest of Canada, a bigger rise in the comparison group than the two averages show. Its estimate was 1.03, with a ninety-five percent confidence interval from 0.83 to 1.29. A confidence interval is the range of effects the data are reasonably compatible with, and this one runs from a seventeen percent fall to a twenty-nine percent rise. The published estimate is the one to quote.

**Sarah:** So the study can't separate a modest fall, no change and a rise. Nguyen and colleagues found hospitalizations fifty-eight percent higher, measured from a baseline before safer supply began in 2020 and compared with a synthetic British Columbia built from six other provinces.

**Kiffer:** Which means their figure combines safer supply, decriminalization and anything else that changed after early 2020. The two studies answer different questions, and reading them together is more useful than picking a winner.

**Sarah:** Here's a harder one, with made-up numbers. After a policy begins, a hypothetical province's overdose death rate rises from twenty to twenty-four per one hundred thousand. Its comparison group rises from five to seven. Did the policy make things better or worse? Take a few seconds.

*(Pause)*

**Kiffer:** On the absolute scale, the province rose by four and the comparison by two, so the difference-in-differences is plus two per one hundred thousand. That looks worse. On the relative scale, the province's rate was multiplied by 1.2 and the comparison's by 1.4. Divide 1.2 by 1.4 and you get about 0.86, which looks fourteen percent better.

**Sarah:** So the same four numbers say worse on one scale and better on the other.

**Kiffer:** Both calculations are correct. They rest on different counterfactuals. The absolute version assumes that, without the policy, the province would have gained the same number of deaths per one hundred thousand as the comparison. The relative version assumes it would have grown by the same proportion. When two places start far apart, those assumptions predict very different things.

**Sarah:** And that isn't only a classroom puzzle. British Columbia's opioid death rate started almost three times as high as the rest of Canada's. How would a researcher decide which scale to trust?

**Kiffer:** One common check is to look at the years before the policy and see on which scale the two places actually moved in parallel. For a reader, the practical habit is simpler. Before quoting a result, find out which scale it's on, and ask whether the other scale would tell the same story.

**Sarah:** Question three is about legal regulation. Is cannabis legalization in Canada working? And give me something to try, because this is where I'd misread a headline.

**Kiffer:** Working on which purpose is the first thing to ask, since the Act lists seven. Let's start with the illegal market, and here's one for everyone. In Health Canada's Canadian Cannabis Survey, among people who had used cannabis in the past year, the share whose usual source was illegal fell from twenty-eight percent in 2018 to three percent in 2024. What does that tell you about everyone aged sixteen and older in Canada? Take a few seconds.

*(Pause)*

**Sarah:** Twenty-eight minus three is twenty-five. So about a quarter of Canadians switched from illegal sources to legal ones.

**Kiffer:** Check the denominator first. Twenty-eight percent of whom?

**Sarah:** Of people who used cannabis in the past year. Most Canadians aren't in that group at all, so I took a share of consumers and read it as a share of everyone.

**Kiffer:** And there's a second slip hiding in it. Saying people switched assumes we followed the same people over time. The published shares compare the mix of usual sources in two different years, and another source shrank too. In 2018 the most common usual source was sharing with friends or family, at sixty percent, and by 2024 it was fifteen percent. And to turn a share of consumers into a share of everyone, you need to know how many people are consumers.

**Sarah:** So I need the share who use. Past-year use among people aged sixteen and older was twenty-two percent in 2018 and twenty-six percent in 2024. Twenty-two percent times twenty-eight percent is about six percent. Twenty-six percent times three percent is about 0.8 percent. So among everyone sixteen and older, the share whose usual source was illegal went from about six in a hundred to fewer than one in a hundred.

**Kiffer:** That's the right reading, and it shows real progress on the Act's purpose of displacing the illegal market. Two cautions. Health Canada notes that people knew the survey was about cannabis, so people who use it may be more likely to take part, and the three percent is rounded, so the 2024 figure is approximate. And the government's expert panel, in March 2024, still judged the illegal market entrenched. The survey asks consumers about their usual source, while the panel looked at illegal production and sale that carry on. Both can be true.

**Sarah:** The other side of the ledger is young children. In Ontario, emergency visits for cannabis exposure among children under ten went from about two and a half a month before legalization to about twenty-two and a half a month after edibles reached stores in 2020.

**Kiffer:** What makes that evidence persuasive about one rule is Quebec, which banned edibles such as candies and desserts. In Ontario, British Columbia and Alberta, which allowed them, hospitalizations for cannabis poisoning in young children rose to 7.5 times the rate before legalization, while in Quebec they stayed at about three times. Quebec works like a comparison group: same country, same federal law, and a different rule on edibles.

**Sarah:** Though Quebec differs from Ontario in plenty of other ways.

**Kiffer:** It does, so the comparison strengthens the link without proving it. It still points to a specific lever, which products may be sold, that a government can change while keeping a legal market.

**Sarah:** Alcohol makes the same point on a much larger scale. The Canadian Centre on Substance Use and Addiction estimated nearly seventy-four thousand deaths from substance use in Canada in 2020, and alcohol and tobacco caused more than eighty-five percent of them. Both are legal.

**Kiffer:** Those figures partly reflect how many people use alcohol and tobacco. They show that a legal, commercial market can coexist with very large harms. The decision to legalize comes first, and then the rules on price, products, marketing and availability decide much of what happens next. HSCI 402 develops that evidence.

**Sarah:** Question four takes us back across the whole course. Is a syndemic just a new word for problems that happen together?

**Kiffer:** It's a stronger claim. Conditions that occur together are comorbidity. Singer and colleagues add two more requirements for a syndemic: social factors, such as poverty, stigma or criminalization, that create the clustering, and an adverse interaction between the conditions that makes the total burden worse.

**Sarah:** Here's one for everyone. Suppose all you knew about Vancouver's Downtown Eastside in the mid-1990s was that public health emergencies had been declared there over overdose deaths and over HIV and hepatitis C among people who inject drugs. Which of the three criteria would that show, and what else would you need before calling it a syndemic? Take a few seconds.

*(Pause)*

**Kiffer:** It shows clustering, the first criterion, and nothing more. For the second, you'd need evidence that social conditions such as poverty, housing and the criminalization of drug use drove the clustering. For the third, you'd need evidence that the conditions made each other worse, for example through the sharing of injection equipment.

**Sarah:** Or through illness and stigma making it harder for people to reach care for either condition. So what does the idea change in practice?

**Kiffer:** It changes what a fair evaluation looks like. Most of the decriminalization studies we discussed counted one or two outcomes, such as deaths or hospitalizations. A syndemic view also asks about police contact, new infections, experiences of stigma and contact with services, and whether effects differed across groups. A policy could do little for one strand and a good deal for another.

**Sarah:** And it's a map of where the series goes next.

**Kiffer:** Right. HSCI 214 develops stigma, trauma, colonialism and mental illness. HSCI 212 develops infections among people who use drugs. HSCI 431 develops the global HIV epidemic and syndemic theory itself, including its critiques.

**Sarah:** And HSCI 402 develops treatment, harm reduction, safer supply and policy evaluation, while HSCI 485 develops psychedelic medicine.

**Kiffer:** Which brings us back to British Columbia's decision, because that's one Sarah and I read a little differently.

**Sarah:** Okay, the part where we disagree. I think British Columbia's decision to end the exemption was defensible on the evidence.

**Kiffer:** Then make the strongest case you can for it.

**Sarah:** Neither 2025 study found a significant change in deaths, and Nguyen's team found opioid hospitalizations fifty-eight percent above their baseline. In Oregon, Spencer estimated extra deaths, and the 2026 working paper estimates more, though it hasn't been peer reviewed. Oregon reversed course after three and a half years, and its services pathway barely ran through the courts. British Columbia narrowed the exemption over public drug use, and the health minister said the pilot "has not delivered the results we hoped for." A government can reasonably decide that a policy hasn't earned renewal.

**Kiffer:** That's a fair case, and part of it I share. I wouldn't tell anyone that decriminalization alone will lower deaths, and the British Columbia evidence doesn't show that it did. Where I'd push back is on what one year can show. Possession offences fell by about two thirds while the exemption applied in full. Gaudreault's team found no significant change in hospitalizations compared with the rest of Canada, and Nguyen's fifty-eight percent combines two policies.

**Sarah:** And Oregon?

**Kiffer:** In Oregon, two of the four analyses we discussed found no significant association, one of them after adjusting for fentanyl's spread. Spencer's peer-reviewed estimate points the other way, and the working paper is still unreviewed. So the evidence on deaths there conflicts. Portugal has kept its approach since 2001, and Hughes and Stevens found no major increase in drug use, with falls in problematic use, new HIV cases among people who use drugs, and the share of prisoners held for drug offences.

**Sarah:** But Portugal paired decriminalization with a big expansion of treatment, and it began in 2001, long before fentanyl dominated the supply here. Hughes and Stevens themselves said they couldn't attribute those changes to decriminalization alone.

**Kiffer:** Agreed, and that cuts both ways. Portugal can't tell us that decriminalization lowers deaths here, and one year in British Columbia can't tell us much either way. My own reading is that the exemption delivered its most direct aim, far fewer possession offences recorded by police while it applied in full, and that the health evidence is short and partly conflicting. I can also see why a government that weights public order heavily would end it.

**Sarah:** I'd put more weight than Kiffer does on the fact that two places tried it and both pulled back. When a policy is reversed twice, that says something about whether it works in practice, even when the studies are mixed.

**Kiffer:** That's a reasonable point. A reversal can reflect public order and community concerns as well as health data, so it's partly a point about values too. Where does that leave us on a rule we can share?

**Sarah:** When anyone calls a decriminalization a success or a failure, ask which outcome they're counting, compared with what, and over how long. Then say which outcomes you weight most, because that's where people who agree on the facts part ways.

**Kiffer:** Let's pull it together with three things to take away.

**Sarah:** First, decriminalization and legal regulation pull different levers. Decriminalization changes what happens to the person holding a small amount, and only a lawful or prescribed supply changes what's in the drug.

**Kiffer:** Second, when evaluations of the same policy disagree, compare their comparison groups, baselines, follow-up periods, scales, and how they handled events like fentanyl's arrival. Absolute and relative differences can even point in opposite directions.

**Sarah:** And third, judge any policy on more than one outcome. Cannabis legalization moved most consumers to legal sources and cut possession charges, while poisonings in young children rose most where edibles were sold. A syndemic view asks what a policy did across every strand it touches.

**Kiffer:** If you'd like more practice, rework the hypothetical difference-in-differences problem from this episode with a comparison group that rises from five to six, and see whether the two scales still disagree.

**Sarah:** This is the last episode of the course. The strands we started in Drugs and Society carry on in HSCI 402, HSCI 214, HSCI 212, HSCI 431 and HSCI 485.

**Kiffer:** Take care, everyone.

**Sarah:** And thank you for listening all the way through the course.
