Changelog
Source:NEWS.md
BayesianQDM 0.1.0
Initial Release
This is the initial release of BayesianQDM, providing comprehensive methods for Bayesian quantitative decision-making in clinical trials.
Core Features
Binary Endpoints
-
BayesDecisionProbBinary()- Calculate Go/NoGo/Gray probabilities for binary outcomes -
BayesPostPredBinary()- Calculate posterior and posterior predictive probabilities -
pBetadiff()- Cumulative distribution function for difference of beta variables -
pBetaBinomdiff()- CDF for difference of beta-binomial variables -
AppellsF1()- Appell’s first hypergeometric function implementation
Continuous Endpoints
-
BayesDecisionProbContinuous()- Calculate Go/NoGo/Gray probabilities for continuous outcomes -
BayesPostPredContinuous()- Calculate posterior and posterior predictive probabilities -
pNIdifft()- Exact numerical integration method for t-distribution differences -
pWSdifft()- Welch-Satterthwaite approximation method -
pMCdifft()- Monte Carlo simulation method -
pINLAdifft()- INLA-based method for external data incorporation
Study Designs
- Controlled design - Standard randomized controlled trials
- Uncontrolled design - Single-arm studies with historical controls
- External control design - Power prior incorporation of historical data
Prior Distributions
- Binary endpoints - Beta priors with flexible parameterization
- Continuous endpoints - Normal-Inverse-Chi-squared conjugate priors and vague priors
Calculation Methods
- NI (Numerical Integration) - Exact calculation using convolution
- WS (Welch-Satterthwaite) - Fast approximation for unequal variances
- MC (Monte Carlo) - Simulation-based flexible approach
- INLA - Integrated Nested Laplace Approximation for external data
Documentation
- Comprehensive function documentation with examples
- Three detailed vignettes:
- Introduction to BayesianQDM
- Binary endpoints analysis
- Continuous endpoints analysis
- Complete test suite using testthat