Purdue Didn't Get Caught Once. It Kept Selling to Diverters for a Decade After.

In late 1995 the FDA approved OxyContin, and on the label it wrote four words that would do more selling than any sales rep ever could: the delayed-absorption formulation was 'believed to reduce' the drug's abuse liability.⁴ It was a regulatory hedge, the kind of soft phrase agencies use when they don't actually know. Purdue read it differently. In an internal 1995 report the company called that label language 'so valuable' it could serve as OxyContin's 'principal selling tool.'⁴ Most companies treat their label as a constraint. Purdue treated it as a weapon — and then aimed it at a market the drug had never been studied to serve.

The story everyone tells is tidy: Purdue oversold a painkiller, addiction spiraled, the company got caught in 2007, paid $600 million, and the chapter closed. Almost the only true part is the $600 million. The far more damning fact is what Purdue did after it got caught the first time — and that it admitted to it under oath.

The crisis was engineered, not stumbled into

Here is the thesis, plainly: Purdue did not lose control of a drug — it built a machine to manufacture demand, and kept the machine running long after a federal court told it to stop. The 2007 guilty plea, entered by a Purdue affiliate, covered misbranding OxyContin as less addictive than rival painkillers and cost the company about $600 million, roughly 90% of the OxyContin profits earned during the offense.³ That looks like a reckoning. It wasn't. In November 2020, Purdue Pharma LP — the parent, this time — pleaded guilty to three federal felony counts covering misconduct from May 2007 through at least March 2017.¹ Read those dates twice. The second crime begins the same year the first plea ended. The 'we didn't know' defense doesn't survive its own calendar.

How a five-sentence letter became a sales script

Every demand machine needs a fuel, and Purdue's was a sentence it didn't even write. In 1980 the New England Journal of Medicine published a five-sentence letter to the editor by Porter and Jick. It studied only hospitalized patients receiving short-term narcotics, it contained no original research data, and it was never meant to say anything about long-term outpatient use. It said it anyway, in the hands of others: cited 608 times through 2017, with 72% of those citations using it as proof that addiction was rare in opioid patients.⁷ The myth that fewer than one in a hundred patients would get hooked predates OxyContin. Purdue's contribution was to systematize it — to take a circulating misreading and put it in the mouths of a sales force calling on doctors. In 2017 the NEJM took the extraordinary step of appending an editor's note to a 37-year-old letter, flagging how badly it had been abused.⁷ A journal does not annotate its own back issues unless something has gone very wrong downstream.

The machine itself was hardware. Purdue's unsealed 83-page launch plan, dated September 27, 1995, describes a 351-person sales force it called 'our most valuable resource' — and, crucially, shows the company already aiming past cancer pain at the vast chronic non-malignant pain market, the one for which the drug had the thinnest evidence.⁵ That sales force was then fed money. Internal records covering 1996–2002 show more than $400 million in promotional spending, roughly three-quarters of it after 2000, funding more than 20,000 pain-related educational programs and over $18 million in medical-journal advertising.⁶ By 2003 some FDA officials said those ads 'grossly overstated' the drug's safety.⁶ The label gave the claim; the letter gave the statistic; the sales force gave it legs. That is the whole engine: a regulatory hedge dressed up as science, repeated into hundreds of thousands of exam rooms.

Wasn't the FDA the real failure here?

The fair objection is that Purdue couldn't have done any of this without a regulator that approved the label, blessed the marketing for years, and was slow to act. There's truth in it. The FDA approved the 'believed to reduce' language in late 1995,⁴ and FDA officials themselves later conceded the journal ads overstated the drug's safety.⁶ But 'the FDA let us' is not the same as 'the FDA fooled us.' Purdue's own internal report named the label phrase as a selling tool before a single pill shipped⁴ — that is intent, not innocence. And the conduct that mattered most came after every warning had landed: Purdue admitted, in 2020, to paying kickbacks through a doctor speakers program, to paying a records company to push opioid prescriptions at the point of care, and to continuing to market to more than 100 providers it had good reason to believe were diverting the drug.⁸¹ A complicit referee does not absolve the player who keeps fouling after the whistle.

Purdue's defenders have always reached for the language of accident — a drug that 'got away from us,' a crisis 'no one could have foreseen.' Its own documents and its own guilty pleas tell a colder story. The crisis didn't get away from Purdue; Purdue built the road it traveled down. A label hedge it privately called a selling tool, a letter it knew said nothing about chronic pain, a sales force it called its most valuable resource, and prescribers it admits it kept paying even as it watched the drug being diverted. The largest criminal penalties ever imposed on a drugmaker² are not the price of a mistake. They are the receipt for a decade of choosing to keep selling — and the proof that the most dangerous product Purdue ever shipped was the belief that the product was safe.