Innovation & Tech

SAS and Posit Connect Tools to Govern AI in Clinical Trials

SAS and Posit will link Viya and Clinical Acceleration with Posit Workbench, putting AI-assisted SAS and R coding inside governed, auditable clinical environments. Regeneron backs the approach.

By Olivia Hart

3 min read

SAS and Posit Integrate Tools for Auditable AI in Clinical Research
SAS and Posit Integrate Tools for Auditable AI in Clinical ResearchKost9n4 / pixabay

What's News

  • SAS and Posit PBC announced their partnership on November 16 to integrate SAS Viya and Clinical Acceleration with Posit Workbench, Positron, Posit Connect and Posit Package Manager.
  • The integration unifies SAS, R, Python and SQL workflows under shared permissions, version control and audit trails.
  • Regeneron's Dilip Raghunathan said earlier mixed R-and-SAS workflows involved workarounds, copied data and duplicate access-control governance.
  • SAS claims the controls support FDA requirements for electronic records and signatures.
  • SAS and Posit have not yet decided which clinical workflows to support after the initial integration.

SAS and Posit PBC announced a partnership on November 16 to connect SAS's clinical-research technology with Posit's open-source development tools, starting with regulated clinical-development workflows that mix SAS and R code. Posit, formerly known as RStudio, supplies the development environment; SAS supplies the analytics, access controls and auditability that regulators expect.

The deal integrates SAS Viya and SAS Clinical Acceleration with four Posit products: Posit Workbench, Positron, Posit Connect and Posit Package Manager. For clinical programmers, that means SAS and R no longer have to live in separate, separately governed worlds.

What does the integration actually change?

Eric Pite, senior vice president of product and strategy at Posit, said a programmer can write SAS and R code in Positron, with Viya executing the SAS portion under that programmer's own permissions. The connected experience also covers Python and SQL alongside the two headline languages.

Customers can work in managed, versioned R environments while controlling dependencies and interoperability constraints through Clinical Acceleration. Posit Package Manager keeps R and Python environments consistent and reproducible, and teams can develop code with Positron and Posit Assistant. In practice, AI-assisted coding now runs inside the same managed setup as the rest of the analysis.

That last point is the crux. AI code generation has spread through clinical programming, but usually outside the controlled environments that submission datasets require. The takeaway: the integration moves AI assistance from the shadow workflow into the audited one.

How does AI assistance stay auditable?

Clinical Acceleration provides the clinical-data layer: a secure repository, a statistical computing environment and AI-assisted transformation tools for preparing research data for regulatory submission. The repository can consolidate information from electronic data capture and clinical data management systems, laboratories and contract research organizations, with traceability back to source data.

For AI-assisted standardization, SAS's documentation describes a workflow that recommends data mappings, reuses transformation assets and generates reviewable SAS code. Human-review workflows and traceable decisions keep experts in control of the output. SAS says AI assistance therefore remains subject to inspection as teams create submission-ready datasets.

Access and compliance controls span roles, permissions, electronic signatures, version control and audit trails. SAS claims they support FDA requirements for electronic records and signatures, while Viya provides the underlying data and AI platform with governance and lineage built into data access. The claim is SAS's own; the partnership announcement does not include independent verification or named customer pilots beyond Regeneron.

In plain words: the vendors promise that an AI-suggested transformation can be traced, reviewed and signed off like any human-written step — provided the controls work as described.

Why Regeneron's workaround history matters

Regeneron's experience illustrates why those controls matter in daily programming. Dilip Raghunathan, a business solutions partner at the company, said earlier mixed R-and-SAS workflows involved workarounds, copied data and duplicate access-control governance.

Programmers now write both languages in one session under their own access rights, with AI assistance operating within those controls. "We can tackle complex multilingual workflows seamlessly without overheads that drew us away from the scientific question," Raghunathan said.

The quote carries weight, but it describes one early customer. The announcement offers no figures on deployment cost, timeline or how many sponsors have signed on.

What comes next?

Beyond the initial integration, Pite said SAS and Posit are still working with joint customers to decide which clinical workflows to support next. No roadmap date accompanied that statement.

The immediate change is a shared framework for language choice, AI assistance, access control and reproducibility. The partnership's wider significance rests on how many regulated clinical workflows the companies bring into that framework — a question neither vendor has yet answered with numbers.

Source: unite.ai

Share this article:

More from Olivia Hart

Olivia Hart

Show full bio

Staff writer covering people, launches and day-to-day practice in entrepreneurship, startups and venture funding at Business Bearings.

2,205 articles

Related articles

« Previous articleNext article »