Implement backfill cohorts simulations

- [x] Adapt `Data` class - [x] `backfilled` and `response` slots - [x] Adapt `update` method for `Data` - [x] adding `response`, `backfill` and `cohort` arguments - [x] Amend `plot` method for `Data` - [x] how to best show response and/or backfill status? - [x] `RandomCohortSize` class / methods - [x] Adapt `StoppingPatientsNearDose` class / method - [x] `Opening` classes and `openCohort` methods: - [x] `OpeningMinDose` - [x] `OpeningMinCohorts` - [x] `OpeningNone` - [x] `OpeningMinResponses` (to use the response efficacy data) - [x] `Recruitment` classes and `maxRecruits` methods: - [x] `RecruitmentUnlimited` - [x] `RecruitmentRatio` - [x] `Backfill` class - [x] tidy methods - [x] `Data` - [x] Adapt `Design` class - [x] add `backfill` slot - [x] Adapt `simulate` method for `Design` class - [x] add `trueResponse` argument - [x] adapt code - try to move things into helper function to keep function size more or less - [x] Consider adapting `simulate` methods for `DualDesign` and `DADesign` or assert that there is no backfilling in place for those - [x] `knit_print` methods for - [x] `Opening` - [x] `Recruitment` - [x] `Backfill` classes - [x] enable again in `Design` - [x] Add a new vignette about backfilling with examples and documentation