R infrastructure for regulated pharmaceutical statistics — reproducibility auditing, Bayesian methodology, electronic audit logging, and row-level data provenance.
Open-source R tooling for regulated clinical analysis — an infrastructure suite under the ReproStats organisation, and a Bayesian clinical trial suite by Ndoh Penn.
The reproducibility foundation. Audit, certify, and monitor R scripts and environments for regulated settings. Designed around 21 CFR Part 11-aligned compliance processes.
Electronic audit logging for R, designed around 21 CFR Part 11 requirements. Tamper-evident, user-attributed, timestamp-verified logging of all analytical actions. Ships with IQ/OQ/PQ qualification scripts.
Row-level data provenance for clinical datasets. Track every transformation from raw SDTM input to analysis-ready ADaM datasets with complete traceability and CDISC Reviewer's Guide-aligned reports.
Structured Bayesian prior elicitation, conflict diagnostics, sensitivity analysis, and regulatory reporting for clinical trials. Aligned with the FDA 2026 draft guidance on Bayesian methods.
Bayesian interim monitoring using predictive probability. Unified futility and efficacy stopping rules across binary, continuous, and time-to-event endpoints. Integrates bayprior priors directly.
Bayesian operating characteristics simulation for clinical trial design. Simulate type I error, power, and expected sample size under user-specified priors and decision thresholds.
Bayesian sensitivity analysis for clinical trials. Prior sensitivity, missing data tipping points, and estimand-aware sensitivity analyses across binary, continuous, and time-to-event endpoints.
Practical guides, methodology notes, and regulatory context for pharma statisticians. See all posts →
How to run the pre-written IQ/OQ/PQ protocols, capture the qualification record, and log the process itself — for 21 CFR Part 11 and EU Annex 11 environments.
End-to-end audit logging for a Phase III RSV vaccine trial — GMT, seroconversion, electronic sign-off, and hash chain verification.
Building ADEFF with row-level provenance — population flags, change from baseline, EASI-75 responder flag, CONSORT table, and subject-level tracing.
How to elicit, document, and justify informative priors for regulatory submissions — aligned with FDA 2026 draft guidance.
These packages are the open-source foundation. Reprosia is the platform built on top — adding environment validation, submission tooling, and expert services for regulated pharma teams.
Explore reprosia.com →Reproducibility in clinical analyses is harder than it should be. R environments go undocumented, package versions drift, and the steps between raw data and submission output are rarely captured in a way that supports independent verification.
ReproStats is an open-source project building the R infrastructure to address this. Packages covering behavioural reproducibility auditing (reproducr), electronic audit logging (regulog), Bayesian prior methodology (bayprior), and row-level data provenance (lineager).
Alongside the infrastructure suite, Ndoh Penn maintains a Bayesian clinical trial suite (bayprior on CRAN, with baymon, bayoc, and baysen in development) aligned with FDA 2026 Bayesian guidance.
ReproStats is maintained by Ndoh Penn, a biostatistician based in Antwerp, Belgium. The enterprise platform built on this foundation is Reprosia.