About BioStatHub
BioStatHub is an independent software project that builds practical, statistically rigorous tools for experimental design and clinical research. Its current focus is Simon's two-stage Phase II clinical trial designs, both conventional and adaptive.
An independent software project
BioStatHub is developed and maintained by an individual developer. It is not a company, and it makes no claim to institutional backing.
The aim is straightforward: give clinical researchers and statisticians tools for experimental design that are correct, transparent and easy to use. At present the project concentrates on Simon's two-stage Phase II designs for single-arm trials, including an adaptive extension that is offered as a preview.
Why BioStatHub?
Designing a Phase II trial means committing to sample sizes and decision boundaries before any patient is enrolled. Many different designs can satisfy the same Type I error and power requirements, and they differ in what they cost: the maximum number of patients, the number expected when the treatment does not work, and the probability of stopping early.
Choosing well therefore depends on comparing operating characteristics and design trade-offs side by side, not on a single number. BioStatHub brings design generation, comparison and evaluation into one workspace, so the candidates can be generated, compared and documented in one place, and the reasoning behind the final choice can be recorded.
The expertise behind BioStatHub
BioStatHub is developed independently by a software engineer with a background in applied statistics and experimental design. The background includes:
- A PhD in applied statistics.
- Research experience in experimental design and design efficiency.
- Extensive experience in mathematical statistics and statistical methodology.
- More than 25 years of software engineering experience.
This combination of doctoral-level statistical research and long professional experience in software engineering informs BioStatHub's approach to building reliable, reproducible and accessible statistical tools.
Statistical rigor and validation
The calculations are exact. Design probabilities are computed from binomial distributions rather than approximated or simulated, and the statistical logic is kept separate from the user interface.
Results are compared independently against established statistical implementations and published reference results wherever they exist. The comparisons that have been completed so far, with their scope, tolerances and any documented discrepancies, are published on the verification pages.
Reproducibility and transparency are working principles: the inputs behind a design are recorded with it, reports state the parameters used, and the verification material is public so it can be checked.
Validation is an ongoing development activity, not a one-time certification. New designs and features are checked as they are added, and any statistical result should still be reviewed against the requirements of your own protocol.
An evolving statistical software platform
Today BioStatHub is a focused tool: it generates and compares Simon's two-stage designs and previews adaptive extensions of them, with reproducible reports.
The longer-term intention is to develop further specialised statistical tools for study design and analysis. Each will be held to the same standards: sound methodology, transparent results and a usable interface.
Get in touch
Feedback from statisticians, clinical researchers and other users is welcome, whether it is a question about a result, a suggestion, or a possible error.
Email support@biostathub.io, or open the BioStatHub application and use the feedback option there.