Skip to content

NSAID adherence #38

@robynstuart

Description

@robynstuart

Problem

Current adherence logic treats all treatments uniformly: non-adherent individuals immediately stop treatment and progress to the next option in the cascade. This doesn't reflect the reality of NSAIDs and TXA, which are use-at-will medications taken during menstruation, not continuous daily treatments like the Pill or hIUD.

Options considered for NSAIDs/TXA non-adherence

  1. Skip adherence checks entirely - No adherence requirement; only non-responders progress to next treatment

  2. Track adherence but don't stop treatment - Monitor adherence for analysis but allow continued use regardless

  3. Stop treatment but allow re-trying - Non-adherent women stop but can be offered the same treatment again without progressing

  4. Natural discontinuation over time - Probabilistic model where non-adherent women gradually stop using over time rather than abrupt cessation

Cascade progression logic

Two approaches for when women advance to the next treatment:

A. Only non-responders progress - Women only advance if treatment is biologically ineffective; non-adherent women don't progress

B. Both non-responders and non-adherent progress - Current behavior where both reasons advance women through cascade

Implemented solution

  • NSAIDs/TXA: Natural discontinuation (option 4) with per-timestep probability of stopping when non-adherent
  • Pill/hIUD: Immediate cessation (current behavior)
  • Cascade progression: Both non-responders and non-adherent advance (option B)

Discussion

Open to feedback on whether this approach adequately captures treatment adherence patterns or if adjustments are needed based on evidence.

Metadata

Metadata

Labels

No labels
No labels

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions