Feature Addition: "Classic" and "Information Theoretic" Decision Rules

[pc494]

Summary Pavel has had fairly clean copies of these decision rules for some time now, I'm hoping to implement them in crmPack (a) because they are a nice addition and (b) as an opportunity to familiarise myself with structure of the code.

References

• O'Quigley J, Pepe M, Fisher L. Continual reassessment method: a practical design for phase 1 clinical trials in cancer. Biometrics. 1990;46(1):33-48
• Mozgunov P, Jaki T. Improving safety of the continual reassessment method via a modified allocation rule.Statistics in Medicine. 2020;39:906–922. https://doi.org/10.1002/sim.8450

Additional Information I've copied over the template as a rough to do list, although I'm not sure how strictly I'll end up following it.

To do:

• [x] find correct locations for the "..." later on in this checklist
• [x] Implement the class in the file RulesClass
• [x] Implement function in the Rules-Method file
• [ ] Add tests
• [ ] Add to pkgdown.yaml

Prototype code: Check NextBestMinDist.txt Check NextBestInfTh.txt NextBestInfTh.txt NextBestMinDist.txt

[pc494]

NextBestMinDist is already implemented (https://github.com/Roche/crmPack/blob/838781e6ef9a5afad10e006b365908b37c96d529/R/Rules-methods.R#L645) so I'll simply check that the implementation in the .txt matches as QA.

UPD: This implementation was fairly recent (late 2021) so I'm going to trust that it was tested at mergetime - moving on to NextBestInf