Biostatistics and data science for clinical research.
Smart Data Analysis and Statistics (SDAS) develops and applies statistical methods to support clinical research and evidence generation from trials and real-world data.
We work with pharmaceutical and biotechnology companies and academic institutions on clinical trial design, comparative effectiveness, evidence synthesis, and prediction modelling. Particular areas of expertise include external controls, Bayesian borrowing, and simulation-based study planning.
We support research teams with study design, statistical analysis plans, data analysis, and interpretation of results. We also contribute to research publications and scientific discussions with regulators and health technology assessment (HTA) bodies.
Our work often involves assessing whether available data can answer a specific research question, combining evidence from different sources, or developing methods where existing approaches are insufficient. We examine the assumptions behind an analysis and assess how they affect the reliability and interpretation of the findings.
- Clinical trial design and analysis: Develop and evaluate study designs, determine sample sizes, and use simulation to assess trial performance under different assumptions.
- External controls and Bayesian borrowing: Assess external data sources, design comparator strategies, and evaluate how informative priors and borrowing affect treatment estimates and trial performance.
- Real-world evidence and comparative effectiveness: Apply causal inference methods, including target trial emulation, to evaluate treatment effects while addressing confounding, missing data, and differences between study populations.
- Evidence synthesis and meta-analysis: Combine evidence through systematic reviews, individual-participant-data and network meta-analysis, and indirect treatment comparisons.
- Prediction models and precision medicine: Develop, validate, and update models for diagnosis, prognosis, risk stratification, and individualised treatment benefit.
- Disease progression and clinical outcomes: Analyse longitudinal outcomes, survival, and patient-reported outcomes, including their relationships with disease progression.
We develop R packages and interactive tools to make statistical methods accessible and analyses reproducible. Selected projects include:
- metamisc — tools for meta-analysis of diagnostic and prognostic research, including prediction-model performance
- precmed — methods for developing and validating models of individualised treatment effects
- SimTOST — simulation-based sample-size planning for bioequivalence studies
- bcts — Bayesian clinical trial simulation and evaluation of borrowing from external controls
Research is a central part of our work. We develop and evaluate statistical methods for clinical trials, real-world evidence, evidence synthesis, and prediction modelling, often in collaboration with academic researchers and clinical teams.
We contribute to funded research projects through methodological development, study design, statistical analyses, and scientific publications. We also supervise and co-supervise MSc and PhD projects.
We contribute to academic teaching and deliver courses and workshops for researchers and statisticians. Our training covers statistical methods and their application in clinical research.
Several of our courses and learning materials on Thinkific are freely available. Our open-source R packages are free to use, with documentation and worked examples to support their application.
SDAS founder Thomas Debray co-edited Comparative Effectiveness and Personalized Medicine Research Using Real-World Data with Tri-Long Nguyen and Robert W. Platt. The book covers methods for studying comparative effectiveness and personalizing treatment using real-world data.
Founded by Thomas Debray, PhD, SDAS brings together statisticians, epidemiologists, and data scientists, alongside scientific advisors and research collaborators.
Our work combines statistical methods development, applied clinical research, teaching, and open-source software. SDAS is based in Utrecht, the Netherlands.
Contact us to discuss statistical support for your study, a research collaboration, or training for your team.