What happened

• Bristol-Myers Squibb Pharmaceuticals Limited made a voluntary admission that overall survival (OS) results in a promotional article were incorrect.
• The material was published online and in the hard copy version of Urology News (v23 (5) July/August 2019) and related to Opdivo (nivolumab) plus Yervoy (ipilimumab) for first-line treatment of adults with intermediate/poor-risk advanced renal cell carcinoma (aRCC).
• The article stated 30-month OS for intermediate/poor-risk RCC patients was 64% (nivolumab + ipilimumab) vs 56% (sunitinib).
• The Panel noted those percentages were for the intention-to-treat (ITT) population (which included the favourable-risk subgroup), not the intermediate/poor-risk population.
• The correct 30-month OS figures for the intermediate/poor-risk population should have been 60% and 47% for the two arms respectively.
• Bristol-Myers Squibb identified the error on 9 July after the online version was published on 8 July; the hard copy journal was already in circulation.
• Immediate actions included arranging withdrawal of the online article and planning a corrected reprint in a subsequent issue with a clear apology and explanation.
• The company described how the error arose during the approval process: data was initially correct at accuracy review, then changed when a missing reference was inserted; the amendment was not appropriately highlighted and was missed at subsequent accuracy and Code review.
• On 12 July, an amended article with corrected data was placed online (with the original job code) and was re-published online before certification; Bristol-Myers Squibb requested its removal.
• Urology News confirmed full removal of the online article on 17 July after technical complexity delayed removal.

Outcome

• As required by Paragraph 5.6 of the Constitution and Procedure, the Director treated the voluntary admission as a complaint and took the matter up with Bristol-Myers Squibb.
• The Panel ruled the information was inaccurate and ruled a breach of Clause 7.2 (as acknowledged by the company).
• The Panel noted the voluntary admission was being taken up in Case AUTH/3286/12/19.

Clauses

• Clause 7.2 (breach ruled by the Panel; inaccurate information in promotional material).
• Clause 14.1 (Bristol-Myers Squibb stated that, through its investigation, it also considered it had breached this clause).

Sanctions

• No explicit additional sanctions stated beyond the required undertaking/corrective actions described in the report.
• Corrective and preventive actions described included removal of the online article, a corrected reprint with a clear apology, retraining and reminders to originators/signatories, ceasing to work with the communications agency involved, engagement with the external accuracy-review vendor, and external audits of approved materials (including an audit of accuracy reviews).

ABPI signatory lens

Why this matters: Promotional claims must be accurate and capable of substantiation. Misstating survival outcomes (even unintentionally) can mislead clinical decision-making and undermines confidence in the certification process, especially when third parties and multiple review rounds are involved.

Where teams typically slip up (interpretation):

• Assuming “minor” edits (eg, adding a reference) cannot change the meaning of data, and therefore not re-checking the surrounding claim.
• Relying on reviewers to “spot the difference” when changes are not clearly tracked/highlighted between versions.
• Allowing publication steps (agency/journal uploads) to proceed before final certified files are locked and verified as the exact version to be published.

The control that would have prevented it: A strict version-control and change-control workflow requiring (1) tracked changes or a change log for every resubmission, (2) mandatory re-accuracy check of any section touched during edits, and (3) a final “golden master” sign-off against the publisher-ready proof before any posting/printing.

What I’d check in an audit:

• Evidence that each review round used clear versioning, with changes highlighted and acknowledged by the accuracy reviewer and Code signatory.
• Whether publisher-ready proofs (online and print) were compared to the certified “final” and signed/dated prior to release.
• Controls for third-party agencies: contractual obligations, training records, and documented escalation when uncertified content is at risk of publication.
• Incident handling: timestamps for withdrawal requests, confirmation of removal, and documentation of corrective communications (eg, reprint/apology distribution).

What the sanctions tell you (interpretation):

• The case outcome emphasizes corrective action and process remediation rather than punitive measures, consistent with a voluntary admission and prompt steps to correct.
• Regulators/panels still expect robust prevention: the narrative focuses on how the error passed through multiple controls, not just the initial mistake.
• Operational risk extends beyond certification: publication mechanics (online posting, digital issue constraints) can materially affect compliance timelines and exposure.

3 questions to ask your team this week:

1. How do we ensure every post-review edit is visible, justified, and re-checked for accuracy before certification?
2. Do we have a documented “publisher-ready proof” step that prevents any online/print release until the exact final file is verified and signed?
3. When working with agencies/journals, who owns the stop/go decision and how do we evidence rapid takedown if something goes live incorrectly?

Key facts

Download the full case report (PDF)

Reviewed by Dr Anzal Qurbain (FFPM) — ABPI Final Signatory