Open shyamvis opened 6 years ago
Great. Look at the dx table and create a histogram of number of patients per IC9/ICD10 code.
Shyam
On Wed, Sep 19, 2018 at 9:13 AM Ke Yu notifications@github.com wrote:
@shyamvis https://github.com/shyamvis I just got the connection to Neptune. I am able to access to SAM_ELT now.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/kayhan-batmanghelich/BatmanLab/issues/64#issuecomment-422798352, or mute the thread https://github.com/notifications/unsubscribe-auth/AIR8_-dkqBXTXdyFLQXkHY-51-R7UITTks5uckMMgaJpZM4Wh4IP .
Top 10 ICD9 codes by #patients (DX_PERSON_CNT is the #patients, DX_PERSON_PROP is the proportion of total patients)
Top 10 ICD10 codes by #patients (same column format)
Observation - the rank of top 10 codes are different between ICD9 and ICD10
Histogram - use log scale because the distributions have very long tail
ICD9
ICD10
There are also some records with unknown ICD type
Great. I will ask Patrick to investigate the unknown coding system data. We are supposed to get it down to a smaller number.
@shyamvis I just got the connection to Neptune. I am able to access to SAM_ELT now.