I fact checked that Atlantic piece on drugs. It's even worse than I expected.

The first time I started reading "The Cities that Said Yes To Drugs," from this month's issue of The Atlantic, I couldn’t make it past the first couple paragraphs— people who use drugs were portrayed as a mass of "zombies" reduced to their bodily functions. I thought, I know where this guy stands, I don’t need to read his whole argument.    

But then I decided to try again. Because plenty of people assume harm reduction is not a scientifically validated strategy, and they might see author Michael Powell’s insistence on calling people who use drugs “addicts,” for example, as evidence not of bias but his unwillingness to soften the truth. So I decided to strip away the rhetoric and see if any skeleton of a legitimate argument remained.

I expected fact-checking a 5600 word piece to be a slog, but I had no idea how big a project this would be. Nearly every paragraph is somehow misleading, and I wondered how I’d sift through it all. But I felt revitalized in my mission when I got to a segment that made me have to peel my jaw up off my desk. This man, who was framing himself as the hard-nosed, fact-slinging foil to all the touchy-feely harm-reductionists had, in a single sentence, made multiple errors in elementary school level math. More detail on that in a moment.

I don’t want to make too much of these mistakes, as remarkable as it is that they’ve gone unchecked since their online publication in July, and have now been emblazoned in ink in this month’s print edition of The Atlantic.  Powell’s main data problem isn’t mathematical—it’s that he picks on specific cities while omitting the national context—to support the misleading idea that harm reduction is somehow contributing to overdose fatalities. Even without the errors, the core deception would remain.

The goal of this piece is not to make a comprehensive policy counter-argument, or to debate whether the cities Powell targets had blind spots in their harm reduction strategy. It is simply to examine the evidence contained within the piece. And I’ve concluded that the best critique of Powell’s argument—is to look closely at Powell’s argument.

Here is a condensed summary of the most dubious elements. Paid subscribers can view a longer, more detailed breakdown below the summary—which anyone can read with a free trial upgrade to the “deep reporting” tier. The full piece is, I’m sorry to say, even longer than Powell’s, but there is plenty of juicy stuff that’s too detailed to include here. I’ve included navigation tools in the full piece, so you can jump to whatever interests you most.

TL;DR 

• On “people who use drugs” versus “addicts:” Powell doesn’t fully explain why he insists on using the term “addict” (which he does 50 times throughout the piece) though he briefly explains that harm reductionists prefer the phrase “people who use drugs,” because it's less stigmatizing. This is true, but “people who use drugs” is also more accurate, because not everyone who uses drugs has a substance use disorder, and, crucially, neither does everyone facing overdose risk.
• Powell singles out cities that embraced harm reduction, and implies these policies failed because recent overdose fatality numbers are much higher than they were in 2014 . It is deeply misleading to attribute these increases in fatalities to harm reduction policies, because the same trend applies nearly universally in the US, and fatality rates are often higher in regions that prioritize criminalization over harm reduction. Public health researchers refer to this trajectory in waves—the earliest wave of the overdose crisis was linked to prescription opioids, the second to heroin, the third to fentanyl, and the fourth to polysubstance overdoses usually involving fentanyl. These waves are not a full explanation of changes in overdose fatality numbers (the pandemic certainly played a role) but fentanyl is more likely to be lethal than other opioids, and crackdowns on heroin and prescription opioids directly contributed to fentanyl's prevalence.
• Powell makes basic multiplication and subtraction errors in his section singling out British Columbia as an example of harm reduction policy failures in Canada. He refers to an “increase over fifteen years” when the period he references is actually several years shorter, and then incorrectly calculates that fatalities increased by a factor of nine. 
• In his screed against progressive drug policies in Canada, Powell relied heavily on conservative publications and political talking points for his facts, even when claims are poorly sourced or flat-out wrong (like his claim that it a BC court blocked a law banning drug use on playgrounds.)
• While Powell admits in a few places that overdose prevention centers (OPCs)—where people inject drugs under professional supervision—do keep people who use drugs safer, he still insinuates that cities who want to establish an OPC are wrongheaded. What he doesn’t explain is the scale of existing OPCs’ success. North American OPCs have overseen millions of injections, reversed tens of thousands of overdoses, with only one reported fatality. Injecting at an OPC is not just safer, it is astronomically so. 
• Powell refers to an unspecified study about an OPC that shut down in Alberta to show that the closure didn't cause harm. But firefighters in the Alberta cities of Calgary and Red Deer—which Powell references without naming—have both complained that overdose related emergency calls have risen dramatically since local OPCs closed.
• Powell writes about “harm reduction” as though it’s an alternative approach to substance use treatment, when it is not—it’s a set of strategies designed to keep people who use drugs as safe as possible until they’re ready for treatment, if that ever happens. While Powell’s framing implies that policymakers who favor harm reduction are somehow in opposition to policymakers that favor treatment, this is far from the case. In fact Burlington, one of the cites singled out as a bad example, has launched multiple low barrier pilots for buprenorphine treatment access. Powell doesn't really define what he means by "harm reduction" or "treatment" in the piece, nor does he ask or answer questions about what type of treatment should be expanded, how it should be paid for, etc.
• Another data point that Powell returns to repeatedly is Burlington’s 2020 decision to reduce the police force by 30%. Powell writes about this change as thought it was part of the city’s move toward harm reduction, when in fact it was the result of successful racial justice advocacy in the wake of George Floyd’s murder. 
• Powell says that “evidence” supports his assertions, for example that “forced treatment” is as effective as voluntary treatment, or that OPCs draw crime, without pointing to where that evidence came from.
• In order to create the menacing atmosphere promised in the “zombie apocalypse” subhed, Powell employs anecdotes from the people of Burlington that appear to be based on speculation, including an anonymous woman’s account of building a “safe room” to protect her children "in case of gunfire.” No actual gunfire is mentioned in the reporting.
• Powell is obsessed with people using drugs on playgrounds, and to read his reporting, you'd think there are none left for the children of Seattle. My close friend with a toddler and a baby lives there. Her kids play at local playgrounds all the time.

Here is an archived link to The Atlantic piece, on the off-chance they update it.