Less Than One Percent: Selling OxyContin to Prescribers

Purdue Pharma launched OxyContin in 1996. It was oxycodone in a controlled-release formulation, and the pharmacological argument for it was reasonable: sustained blood levels rather than peaks and troughs, meaning better pain control and — Purdue's claim — a lower risk of addiction, because the euphoric spike associated with abuse liability was smoothed out.

The company's sales representatives told physicians that the rate of addiction in patients treated with opioids for pain was less than one percent.

That figure came from a letter. In January 1980 the New England Journal of Medicine published a five-sentence letter to the editor from Jane Porter and Hershel Jick, reporting that among nearly twelve thousand hospitalized patients who had received at least one narcotic, only four cases of documented addiction were found in patients with no history of addiction.

It was a paragraph about inpatients receiving supervised doses in hospital. It was not a study of outpatient chronic pain treatment, it had no follow-up, and its author later said publicly that he had never imagined it would be used the way it was. A 2017 analysis in the same journal traced over six hundred citations of it, most characterizing it as evidence that addiction was rare in patients treated with opioids.

Purdue's marketing was extensive: a large expansion of the sales force, a database identifying high-volume prescribers, bonus structures tied to volume, sponsored continuing medical education, funded pain-advocacy organizations, and paid speakers.

Three mechanisms operate.

Anchoring. A precise-sounding number delivered by a trusted source establishes a reference point for risk that is very difficult to displace, and clinicians' own priors — which would have been more cautious — were overwritten by a figure with an apparent citation behind it.

Normative influence. Sponsored education and opinion leaders create a visible professional consensus, and Deutsch and Gerard's work establishes that a clinician who is more cautious than their peers experiences that caution as deviance.

And the harm falls on people with no exposure to the evidence at all. The patient receives a prescription; they cannot evaluate the base rate, and they have no reason to.

Purdue and three executives pleaded guilty to misbranding in May 2007 and paid around six hundred and thirty-four million dollars. The company pleaded guilty to further federal charges in 2020.