I would like to use the skills I learnt on my EBHC MSc to develop a prediction type tool to help mothers and their health care team identify more accurately the cause of their pain on breastfeeding.
This would reduce overdiagnosis because everyone leaps to the best known / worst case conditions, not the most probable. Thinking particularly of thrush treatments when the pain is attachment, or Raynaulds or vasospasm. Or confusion between mastitis and blocked ducts when antibiotics are not often necessary because there is no infection. This would involve building more robust disease definitions and gaining consensus - which might be very difficult.
The gain would be in reducing GP and A&E visits and more accurate medicines for those that need it and more importantly reducing emotional toil on new mothers.
How could I do this in a way that reduced bias as much as possible?
I could interview women with pain meeting a set criteria, I think I should try and find outliers which don't meet expected pain patterns and try to challenge the prediction tool.
The complication is that I have no access to a lab to test bacterial or fungal infections. I am a peer supporter with a charity so no link with the NHS either.
Phyll
|