Stop Pitching R. Start Pitching the Model.

August 4, 2026

By Aga Rasinska, Director of Strategy, & Ashley Tarasiewicz, Director, Head of OpenVal^®

Two years ago, pharma companies were still debating whether R belonged in a regulated environment. That debate is over. In the 2025 Pharmaverse R Adoption Survey, 90% of sponsor companies expect their R usage to increase over the next two years – up marginally from 87% in 2024, and the number that expect it to decrease is still zero. Nobody in this sample thinks R is going away.

The technology question is settled. What’s left is harder, and it’s the question the data answers: can organizations run R inside a regulated, auditable, governed operating model, not just an exploratory space?

The 2025 survey, run by Pharmaverse, collected 31 responses from pharma and biotech sponsors (CROs and vendors excluded by design). Read against last year’s baseline, it shows an industry is graduating from “can we use R here” to “how do we run this at scale.”

Exploratory Work Is Turning Into QC

In ADaM, QC usage jumped from 43.5% of companies in 2024 to 64.5% in 2025, the single largest move in the dataset. In SDTM, QC usage more than doubled, from 13% to 32.3%. TLG/TLF work, already broad in 2024, held steady rather than exploded. It was already mature.

That pattern is the story. When a technology moves from “exploratory” to “QC”, it means it’s become trustworthy enough to check other people’s work against. And this shift is happening fastest exactly where the regulatory stakes are highest: ADaM and, more slowly, SDTM.

SDTM is still the laggard: 45.2% of companies report no current use but plans to start, and only 6.5% use R for first-line SDTM programming. But even there, the direction is unmistakable; QC usage in SDTM nearly tripled year over year.

The Obstacle List Hasn’t Changed

Here’s the part that should give every team pause: the top blockers in 2025 are almost identical to what you’d expect from any maturing regulated-industry rollout, and they are not technical.

Resistance to change from existing workflows was the single largest obstacle by a wide margin, with 61.3% of respondents reporting this concern. Package and dependency management followed (48.4%), then lack of internal R expertise (38.7%) in a tie with regulatory concerns and validation requirements (also 38.7%).

Meanwhile, the success factors respondents report included strong internal champions (48.4%), executive sponsorship (45.2%), dedicated training programs (41.9%), and, notably, regulatory precedent from other companies (35.5%).

Put those two lists side by side and the conclusion writes itself. Nobody is stuck because R can’t do the job. Organizations are stuck because change management, dependency governance, and validation strategy haven’t caught up to what the tooling can already do. The companies pulling ahead are the ones who treated foundations as first-class work from day one, instead of something to clean up after the pilot succeeded.

Python Is Showing Up

This year’s survey added Python questions for the first time, and the answer is clear even from a small sample: Python is present, but it’s living in the sandbox. Of 19 written responses describing Python usage, the dominant themes were exploratory work, automation and pipeline tooling, and early ML experimentation, not core regulatory submission work. Only one respondent (out of 31) reported using either of Pharmaverse’s dedicated Python tools, py-pkglite or rtflite.

Python is a second language slowly showing up in the building, but it hasn’t been handed the keys. Whether it earns them, or organizations simply keep deepening their R investment instead, is still an open question.

Submission Is No Longer Theoretical, But It’s Still Rare

Three companies now report doing end-to-end R-based eSubmissions, unchanged from 2024. However, 16 out of 31 say they’re not doing it yet but plan to, and open-text responses describe real timelines and internal readiness programs, not vague ambition. Submission-ready environments came up repeatedly as the top unmet need in this year’s open comments, ahead of metadata handling and package discoverability.

The Read for Anyone Planning the Next Two Years

Team sizes tell the same story from a different angle: 12.9% of companies now report more than 500 R users doing clinical reporting work, up from 8.7% last year, while the “under 50 users” segment shrank from 56.5% to 51.6%. The center of gravity is moving toward scale.

None of this means the hard part is over. It means the hard part changed. A year ago, the question that was open focused on whether R could do regulated clinical work. Now it’s whether your organization has the internal champions, the executive backing, the training pipeline, and the discipline to run it as workflow rather than a side project.

If you’re the one building the business case internally: stop pitching R. Pitch the model. That’s what convinces leadership: a model of work that leverages open-source tooling, an agile mindset, and collaborative ways of working, not another package demo.

Full survey results and methodology: “How R We Now? Tracking Industry Progress in Adopting R and Open-Source for Clinical Reporting,” Białaś & Tarasiewicz, presented at PHUSE US Connect 2026 in Austin. View the full paper or slides.