misc changes todo

[alistairewj]

• [ ] consider dowhy / match without repetition
  • does this allow for IPW easily
• [ ] run cumulative hazards on matched cohort