Project Planning

Propose to split the work into four streams. For each stream, we need a direct response person to ensure the work is delivered as planned with agreed priority.

• Package development: the goal is to ensure R packages API are user-friendly, and results are accurate.
• Computational efficiency: the goal is to ensure a more efficient version of implementation with accurate results.
• Shiny: the goal is to use API to build Shiny app for general user.
• Publishing: the goal is to provide clear documentation.

Package development

gsDesign2

(Yujie -> Maintainer)

• Software engineering best practice (Yilong).
• Documentation of info_scale (Yujie).
• Examples of updated design (Yujie).
• Vignette of the information under the null and alternative hypothesis (Yujie).
• Add gs_power_ve() and gs_design_ve() (Keaven).
• Add as_r2rtf() and as_flextable() (Yilong with SP).
• Implement stratified sample size method from Bernstein and Lagakos (?).

simtrial

(Nan -> Maintainer)

• add RMST. (?)
• correlated endpoints (Hongtao)

Grammar of design/simulations

(De-prioritize)

• distsurv (Yujie and Nan).
• gscut (Yujie).
• gsspending (Yujie).
• wlrweight (Yujie).
• tidygsd (Yujie)

Computational efficiency

(Nan drive with SP and contractor)

• simtrial -> data.table rewrite (Nan drive with the contractor).
• gsDesign2::gs_bound_spending() -> C++ rewrite (Nan drive with SP).
• gsDesign2::expected_event() -> data.table rewrite (?).

Shiny

• gsDesign2Shiny (Gregory/Bing).
• gsDesignShiny (Nan).

Publishing

• gsDesign2 under NPH paper (Yujie/Yilong).
• gsDesign Shiny book (Keaven/Nan).

Reasearch

• futility bound under NPH
• update bound at the time of analysis (reference: https://journals.sagepub.com/doi/full/10.1177/17407745211072848)