Project
A defined deliverable with a fixed timeline and price: a qualification pack, a reproducibility audit, a migration check.
Best for a clear, bounded needConsultancy for regulated statistical analysis.
Specialising in R and reproducible workflows, including moves from SAS. Reprosia takes on project work, advisory sessions and contract engagements for pharma, biotech and research teams.
Ways to work together.
Pick the shape that matches the need. Most start as a short conversation or a small fixed-scope project.
Contract and freelance
Statistical programming and biostatistics support inside your team, for a few weeks or several months. Remote, or on site by arrangement.
Availability varies: ask for current capacityAdvisory
A second opinion on a validation approach, an SOP set, an analysis review or an AI-use policy, in half-day or day blocks.
Best when you have a team but need a viewSupport retainer
A standing arrangement for questions, reviews and priority fixes around the R environment and the open-source tools.
Best once a project has shippedWhat we help with.
- Regulatory R implementation
- Can we explain our use of R to an auditor?Validation strategy, SOPs and system validation documentation for using R in GxP environments.
- Package and environment validation
- Can we show this package is installed and behaves as specified?IQ/OQ/PQ documentation for R and R packages, using regulog's qualification scripts where they help.
- Reproducibility and provenance audits
- Can someone else re-run this and get the same numbers?An independent review of an analysis pipeline: undocumented dependencies, environment risks, excluded records and a prioritised fix list.
- SAS-to-R migration
- Do the R outputs match what SAS gave us?Equivalence testing between SAS and R outputs, parallel validation documentation and a migration audit trail.
- Statistical programming and analysis support
- Do we have the capacity, and a second pair of eyes, for this analysis?Analysis programming and code review (R specialism, SAS and Python workflows understood), and general statistical consultancy on study analysis, including Bayesian prior documentation with bayprior.
- Shiny application development
- Can our team explore and review results without writing code?R Shiny apps for analysis review, data exploration and internal tools, built with reproducibility in mind: version-controlled, tested, with the environment locked. Whether an app needs formal validation depends on how your team uses it, and that is agreed with your quality group. One public example: the live demo of the bayprior app, an open-source tool for documenting Bayesian priors.
- AI-use governance
- What do we log when an AI assistant helps, and who reviews it?What to log when an AI assistant helps with an analysis, who reviews it and how to replay it. Starts with the free evidence kit.
- Evaluation of AI coding agents Pilot
- How often is the code an AI agent writes actually right?Measuring how often an AI coding agent gets routine programming tasks right, using tasks built from your own specifications: repeated attempts, failures sorted by type, an audit log and a short report for your QA group. A measurement, not a certification, and offered as a pilot. How it works.
- AI-assisted independent QC Pilot
- Who decides when two programs disagree?An AI coding agent writes the second, independent program from your specification; a script lists every difference; your QC lead adjudicates each one and the decisions are logged. Run next to your existing QC on one deliverable, so you can compare. Differences often point to gaps in the specification as much as to bugs. How it works.
Training.
Two topics, taught hands-on with your own kind of data and workflows. Half-day and full-day formats, for statisticians and statistical programmers.
R for clinical statistics teams
Getting productive in R for analysis and reporting, from data handling to tables, figures and reusable code.
Reproducible workflows
Scripted pipelines, version control, locked environments and documented runs, so results can be re-run and reviewed.
How an engagement runs.
Conversation
A short call about your environment, your quality system and what a reviewer would ask. We walk one real workflow with a one-page intake sheet, which you can also use on your own.
Scoping
A short written scope with deliverables, timing and what you need to provide.
Delivery
Documents and code you own, plus the open-source tools where they help.
Follow-up
A review after delivery to deal with questions from QA or auditors.
Start with a conversation.
Describe the need in a few lines, whether it is a project, a contract role, a review or training.