Needs improvement
Sample text on model card: subarachnoid hemorrhage scalp laceration service: surgery major surgical or invasive
Returns the following ICD10 codes:
H59.11 Keratopathy (bullous aphakic) following cataract surgery, right eye
H47.02 Hemorrhage in optic nerve sheath, right eye
M05.75 Rheumatoid arthritis with rheumatoid factor of right hip without organ or systems involvement
I97.12 Postprocedural cardiac arrest following cardiac surgery
S05.01 Injury of conjunctiva and corneal abrasion without foreign body, right eye, initial encounter
Returns similar inaccurate results for "liver cancer"
J18.1
K75.0
L81.1
I24.1
H44.31
Is this the best icd10 predictor model on huggingface, or is there another?
Google AI Gemini with a corpus of docs. Corpus includes CMS descriptions of all ICD10 codes and descriptions
Works quite well even for layman input of symptoms, e.g. high blood pressure = hypertension