In clinical trials, SAS is still the main language for building CDISC datasets and TLFs at sponsors and CROs. R is growing fast: pharma companies share open-source R packages through pharmaverse, and R Consortium pilots have sent R-based submissions to FDA for review. Python is used mainly around trials, for data science, automation and machine learning. FDA does not require any particular software, only reliable, documented work in its data standards. For a first job in clinical programming in India, learn SAS first, then add R.
Key takeaways
- FDA does not require SAS — it requires standard data and reliable, documented software.
- SAS is still the everyday tool for SDTM, ADaM and TLFs at most sponsors and CROs.
- R is the fastest-growing alternative, backed by pharmaverse packages and FDA submission pilots.
- Python is strong for data science, automation and AI around trials, less so for submission work.
- Learning order for a clinical programming job: SAS first, then R, then Python if you need it.
On this page ⌄
Does FDA require SAS? What regulators actually ask for
No. FDA’s Statistical Software Clarifying Statement says the agency does not require any specific software for statistical analyses. What it does require is that the software a sponsor uses is reliable, that its testing is documented, and that submitted data follow FDA’s data standards — CDISC SDTM and ADaM, sent with Define-XML. Sponsors are encouraged to discuss their choice of software with FDA reviewers early.
So the question is not “Which language is allowed?” but “Which language can my company validate, maintain and staff?” That is why the answer differs between companies. For the data standards themselves, read SDTM and ADaM explained.
SAS vs R vs Python: side-by-side comparison
SAS
- Cost
- Commercial licence (free SAS OnDemand for Academics for learning)
- Main use in trials
- SDTM, ADaM, TLFs and submission packages
- Submission track record
- Long — the industry’s default for decades
- Clinical toolkits
- Base SAS, macros, PROC REPORT, SAS/STAT
- Validation
- Validated environments widely set up in pharma
- Learning curve
- Readable, procedure-based
R
- Cost
- Free, open source
- Main use in trials
- TLFs, ADaM, statistics, interactive review apps
- Submission track record
- Growing — R Consortium pilots reviewed by FDA
- Clinical toolkits
- pharmaverse: admiral, sdtm.oak, rtables, Tplyr, xportr
- Validation
- Package validation needed per company
- Learning curve
- Steeper; many ways to do one task
Python
- Cost
- Free, open source
- Main use in trials
- Data science, automation, machine learning
- Submission track record
- Limited public examples so far
- Clinical toolkits
- General data libraries; fewer clinical-specific tools
- Validation
- Package validation needed per company
- Learning curve
- Clean syntax; general-purpose
Why SAS still leads in clinical trials
- Validated code bases: sponsors and CROs have years of tested SAS macros and standard programs for SDTM, ADaM and TLFs.
- Familiar to reviewers: SAS Transport (XPT) files and SAS-based programs have been the norm in submissions for a long time.
- Hiring: clinical programmer job descriptions commonly list SAS and CDISC as core skills.
- Certification: SAS offers formal exams that employers recognise; see clinical SAS certification.
If you are comparing general SAS training with clinical SAS, read SAS vs clinical SAS.
R in clinical trials: how fast is it growing?
R has moved from “for statisticians only” to a real option for submission work:
- pharmaverse is a network of pharma companies and individuals that shares open-source R packages for clinical reporting, such as admiral for ADaM, sdtm.oak for SDTM, rtables and Tplyr for tables and xportr for submission files. Contributors include large pharma companies such as GSK, Novo Nordisk, Merck, Genentech and Biogen.
- R Consortium submission pilots have sent test submissions to FDA since 2021. In 2024 the R Consortium reported that FDA reviewed a pilot with ADaM datasets and TLFs built in R. Later pilots test web-based review apps, Dataset-JSON and AI-assisted analysis.
- Interactive review: R Shiny apps let teams and reviewers explore trial results.
The trend is towards SAS and R side by side, not one replacing the other overnight. Many teams still produce the official outputs in SAS and use R for checks, graphics and new projects.
Python in clinical research
Python is a popular general-purpose language, and in clinical research it is used mainly around the core submission work:
- Automating data checks, file handling and reporting pipelines.
- Machine learning and AI projects, such as predicting enrolment or screening documents.
- Real-world data and data engineering work.
- Reading and writing new formats such as CDISC Dataset-JSON, which FDA is testing as a possible replacement for XPT.
For a pure clinical programmer role building SDTM, ADaM and TLFs, Python is rarely the main requirement today. It becomes valuable if you move towards data science or automation.
SAS vs R vs Python: which should you learn first?
Learn first
- First job as a clinical SAS / statistical programmer at a CRO or sponsor
- SAS + CDISC (SDTM, ADaM, TLFs)
- Biostatistics or statistical analysis
- SAS or R
- Clinical data science, AI or automation
- Python
- Already working in SAS and want to grow
- R (pharmaverse)
Then add
- First job as a clinical SAS / statistical programmer at a CRO or sponsor
- R
- Biostatistics or statistical analysis
- The other one
- Clinical data science, AI or automation
- SAS or R basics + CDISC
- Already working in SAS and want to grow
- Python for automation
Whatever you choose, the skills that carry over are the same: CDISC standards, an understanding of trials and the statistical analysis plan, careful data logic and validation. Learn those once and switching language is far easier. Free practice: SAS OnDemand for Academics for SAS, and R and Python can be installed free — see what is clinical SAS for a starting plan.
Start with clinical SAS at BCRI
BCRI’s Clinical SAS course online is a 3 months live online programme (₹38,000) covering Base and Advanced SAS, CDISC SDTM and ADaM, TLFs, Define-XML and Pinnacle 21, with recordings and placement assistance.
FAQs
Is SAS or R better for clinical trials?
For a first clinical programming job, SAS is still the safer choice because most sponsors and CROs run their submission work on validated SAS code. R is growing fast and is a strong second skill.
Does FDA accept R or Python?
FDA does not require any specific software. It requires data in its standards (CDISC SDTM and ADaM with Define-XML) and reliable, documented software. R-based test submissions from the R Consortium pilots have been reviewed by FDA.
Will R replace SAS in pharma?
Not overnight. Many companies now use SAS and R side by side, and pharmaverse packages make R more practical for submission work. Knowing both makes you more employable.
Is Python used in clinical trials?
Yes, mainly for data science, automation, machine learning and real-world data work. It is less common for building SDTM, ADaM and TLFs for submissions.
What is pharmaverse?
A network of pharma companies and individuals that shares open-source R packages for clinical reporting, such as admiral for ADaM datasets and rtables or Tplyr for tables.
Can I learn SAS for free?
Yes. SAS OnDemand for Academics is free and runs in a browser for learners. R and Python are free to install.