@docarion_hq@silverpenydr@evidencechecked@Doctware Considering Nigeria as a whole, there will definitely be variation in the data, which might introduce inherent biases that need to be addressed.
Maybe these aren’t the questions for now but really need to consider them for deployment sake after your hackathon. I wish you success
@docarion_hq@silverpenydr@evidencechecked@Doctware If it’s the latter, did you consider their retrieval processes or how they were labeled (whether algorithmically or through a manual chart review)? And if former, then was patient consent considered and did it go through proper de-identification process?
@silverpenydr@evidencechecked@Doctware@docarion_hq Finally, how does your model handle the data — Are your models locally hosted? And will data received be used to retrain/finetune the models? Would love to hear more about it pls
@silverpenydr@evidencechecked@Doctware@docarion_hq Great work by your team! I’m very curious, how did you handle the medical text similarity considering the fact that our local languages have low representation in clinical NLP data? Also, how do the doctors get the patients’ medical history to check for contraindications?
@AskMichaelTaiwo I normally wouldn’t volunteer to read literature in class, but that day the teacher chose me herself. Everything was fine until I saw the word “fable”. My brain switched to Yorùbá mode & I confidently said “fàábú. The whole class laughed & “Fàábú” became my nickname for that day.