Research Integrity & Clinical Data — Biotech Scenario

A Phase 3 Trial Shows Positive Primary Results But Disappointing Secondary Endpoints. A Senior Scientist Suggests De-Emphasizing the Secondary Data. Is That Acceptable?

A real biotech and pharma compliance scenario — with three decision options and the right answer.

The Situation

You are a clinical data manager at a biotech company finishing a pivotal Phase 3 trial. The primary endpoint data shows a modest but statistically significant benefit that supports the regulatory submission. However, a secondary endpoint that was pre-specified in the statistical analysis plan — and that the trial was specifically designed to evaluate — shows no meaningful benefit. A senior scientist suggests “de-emphasizing” the secondary endpoint results in the primary publication, noting that “the statistical analysis plan will be on file” and “no one reads secondary endpoints anyway.” You are a listed author on the publication.

Choice A Agree to the framing. The primary endpoint is what regulators and prescribers care about. The secondary endpoint is technically reported in the statistical analysis plan on file. Nothing is being hidden — it’s just a matter of emphasis in a journal article, which is standard scientific writing practice.

Choice B Insist that the pre-specified statistical analysis plan be followed and all endpoints reported with equal transparency. Selective de-emphasis of a pre-specified endpoint is a form of publication bias that violates research integrity standards and, as a listed author, creates personal scientific and legal exposure.

Choice C Refer the decision to the publication committee and ask them to make the call — this is a scientific judgment that should be made by the appropriate governance body rather than individual scientists on the team.

The Right Call

Choice B — Insist on full transparency of pre-specified endpoints. Choice C escalates appropriately but the publication committee must also follow research integrity standards.

Selective de-emphasis of pre-specified endpoints is a recognized form of data manipulation in clinical research — not because anything is falsified, but because the scientific record created by the publication will mislead prescribers, regulators, and future researchers about the drug’s actual clinical profile. As a listed author, you bear personal scientific responsibility for the accuracy of what the publication represents. “The data is on file” does not satisfy that responsibility if the published article systematically minimizes findings that the trial was designed to detect.

Nothing is being falsified — and that makes the violation harder to recognize.

The most obvious form of research fraud is data fabrication or falsification. This scenario involves neither. All the data exists, will be reported somewhere, and the statistical analysis plan is on file. But research integrity standards — embodied in Good Clinical Practice (GCP), the ICMJE authorship guidelines, and most institutional policies — require that pre-specified analyses be reported with transparency proportionate to their pre-specification. Burying a negative secondary endpoint because it complicates the narrative is publication bias, whether or not the data technically exists somewhere.

Listed authorship means personal accountability for what the paper represents.

Journal authorship carries explicit ethical responsibilities under ICMJE guidelines, including responsibility for the integrity of the entire work. A data manager who is listed as an author on a publication that selectively minimizes pre-specified endpoints has lent their name to a document that does not fully represent the data. That is a personal issue of scientific integrity — not just an organizational one.

The downstream consequences affect patients and regulators, not just scientists.

Prescribers who read a publication about a new therapy rely on complete reporting of prespecified endpoints to make clinical decisions. Regulators reviewing the submission require the same. A publication that systematically de-emphasizes a negative secondary endpoint creates a record that the scientific community, clinicians, and patients will act on — often for years. The impact of publication bias in pharmaceutical research is well documented and has contributed to drugs remaining on the market longer than the evidence supports.

What is publication bias in clinical research?

Publication bias is the tendency for clinical research publications to over-represent positive or favorable results and under-represent neutral or negative findings. At the individual study level, it includes selective reporting of endpoints — emphasizing those that show benefit while minimizing or omitting those that do not. It is a recognized form of research misconduct when it involves pre-specified endpoints.

What is a pre-specified statistical analysis plan and why does it matter?

A statistical analysis plan (SAP) is a document created before a clinical trial begins that specifies exactly which outcomes will be measured and how they will be analyzed. Pre-specification prevents researchers from selectively reporting outcomes after seeing the data — a practice called “outcome switching” or “HARKing” (Hypothesizing After Results are Known). When a SAP specifies a secondary endpoint, that endpoint carries an obligation of full transparent reporting regardless of the results.

What are the consequences for a researcher who participates in selective endpoint reporting?

Consequences can include journal retraction of the publication, institutional investigation and sanctions, loss of research funding and clinical trial approval, personal reputational damage in the scientific community, and in severe cases, regulatory action. FDA can also reject or question submissions that are based on publications found to have engaged in selective reporting.

Is it appropriate for a clinical data manager to push back on a senior scientist’s publication decisions?

Yes — and most pharmaceutical companies’ research integrity policies explicitly require it. If a listed author has concerns about the scientific integrity of a publication, they are obligated to raise those concerns through the publication committee, the Chief Medical Officer, or the research integrity compliance function. Seniority of the scientist making the suggestion does not change the reporting obligation.

How to Use This Scenario in Training

Research integrity and Code of Conduct training establishes the policy. This scenario makes it stick.

This scenario is most valuable for clinical operations, biostatistics, medical writing, and medical affairs teams — the people most likely to be involved in publication decisions. The recognition skill is identifying selective presentation of pre-specified data as a compliance trigger, not a scientific writing decision.

More Compliance Scenarios

Compliance Training Built for Biotech and Pharma

Xcelus develops scenario-based compliance training for pharmaceutical and biotech organizations — including research integrity, clinical data ethics, and FDA compliance scenarios built for the specific pressures your clinical and scientific teams face.

View Compliance Programs →
Contact Xcelus