Your Launch Team Is Misreading Your Own Data.

One Idea Worth Acting On.

A novel oncology drug receives full FDA approval based on a 100-patient study.

The launch team takes the data to market.

The medical community immediately calls the evidence weak.

Adoption stalls.

The data was not weak. The data was extraordinary.

Full FDA approval on 100 patients means the signal was so large and so clean that the agency required no more evidence. That is not a small study with weak data. That is an exceptionally strong result that happened to require a small study.

The launch team read it the wrong way. And by the time the statisticians who understood it were finally brought in, the market had already made up its mind.

What's Actually Happening

Biostatistics is still treated as a bookend function. Called at the start to calculate sample size. Called at the end to generate a P value. Invisible in between.

Launch messaging, scientific communication platforms, and field materials are built from top-line clinical results by teams who know how to communicate science but may not understand the statistical architecture beneath it.

The result is organizations sending commercial and medical teams into the market to defend data they have not been equipped to explain — and sometimes to misrepresent data they have not been equipped to read.

As Dr. Matthew Guo, VP of Biostatistics and contributor to Voices of Oncology, describes it: the organization had been receiving the same external criticism for quite some time before anyone thought to bring in the people who could explain what the evidence actually meant. By then it was too late.

Why This Matters

The complexity of oncology data is increasing. Adaptive trial designs, biomarker-defined subgroups, real-world evidence integrations, Bayesian statistical frameworks — all of it requires interpretation that goes beyond the headline hazard ratio.

When the interpretation step is skipped, organizations lose control of their own clinical narrative. They miss strengths they could have defended. They fail to contextualize limitations before critics do. They hand the narrative to the oncology community and hope the community reads it correctly.

It often does not.

The same failure appears in subgroup analyses. When the FDA questioned why a therapy appeared to underperform in the US patient subgroup, the answer was not that the drug worked differently in the US. The answer was that the US sample size was too small to produce a precise estimate — and what looked like underperformance was statistical uncertainty, not biology.

Without a statistician in the room when that question was asked, the wrong narrative almost became the accepted one.

What Needs to Change

Statistical expertise belongs in the launch planning room, the scientific communications process, and the commercial strategy conversation — not just in the analysis plan.

Organizations need people who can generate the numbers and people who can explain what those numbers mean for the specific audience receiving them. Those are different skills. Both are required.

The companies that will control their clinical narrative are the ones that understand their data precisely enough to defend it credibly — before the critics define it for them.

The Bottom Line

Generating strong data is not enough. If the people communicating it do not understand what it means — statistically, not just clinically — the strength of the data does not reach the prescriber, the payer, or the guideline committee.

The launch team that misread the 100-patient approval did not have bad intentions. They had incomplete expertise. That is a structural problem, not a personal one. And it is fixable.

Add Your Voice — Where are you seeing this? What's working — and what's not?