A carregar...
Publicação
Implementation of SAS-based bioavailability assessment in R for phase 1 clinical trials
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 3.12 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
Pharmaceutical companies are increasingly interested in adopting R as an alternative to SAS for clinical trial analyses. While SAS has long been the industry standard for regulatory submissions, R offers advantages such as being open source and providing modern statistical and data visualization capabilities. Before R can be considered in regulatory contexts, the reliability of its results must be demonstrated. In this project, reproducibility with SAS served as a practical way to assess reliability, since SAS is widely accepted by regulatory authorities and generally assumed to produce correct results. The project investigates the use of R for the assessment of relative bioavailability in phase 1 healthy volunteer clinical trials, focusing on fixed-effects and mixed-effects ANOVA models. Analyses were conducted across the main study designs used in practice (crossover, fixed-sequence, and parallel group) using both SAS and R implementations. A simulated dataset was developed and validated to ensure reproducibility, while real case studies were used to evaluate performance in practical settings. The findings show that R can reproduce SAS results under most scenarios. Fixed-effects models yield identical estimates and confidence intervals across platforms, and mixed-effects models show perfect alignment when datasets are balanced. Minor discrepancies in the Kenward–Roger degrees of freedom may arise with incomplete or unbalanced data, resulting in small differences in confidence intervals (at the fourth decimal place). Overall, this work demonstrates that R can serve as a credible alternative to SAS for relative bioavailability analyses, while also highlighting areas for further development. It provides a practical foundation for broader adoption of R in regulatory settings and a launchpad for extending comparisons to additional pharmacokinetic applications.
Descrição
Trabalho de projeto de mestrado, Bioestatística, 2025, Universidade de Lisboa, Faculdade de Ciências
Palavras-chave
R SAS Phase 1 clinical trials Relative bioavailability ANOVA models