Second Opinion - NLP
Problem:
- Unnecessary treatment and malpractice
- Doctors diagnosis remain uncollaborated
- Lack of reliable rating system
Solution:
Web based application that confirms doctors decisions and recommends solutions and an agreement percentage utilizing NLP to analyze hundreds of thousands of other EMRs.
Potential Impact:
Doctor/Nurse verification Patient confidence in diagnosis/treatment -Eliminates need for second opinion
$538M amount 2,588 hospitals will have deducted from their Medicare payments due to high readmission rates
Benefits:
Reduces costs by:
-
Preventing over-treatment
-
Minimizing malpractice
-
Decreasing 'second opinion' visits
-
Supplemental advice
-
Reliable rating system for Doctors
Consumers have more confidence that doctors made the right decision
Future Implementation: Machine Learning for handwriting Doctors can write or speak directly into application Device takes picture for input
Technologies Utilized: TFIDF, ML, wordnet, NLP, XGBoost, Bokeh, Django, synsets, and jupyter