Type 2 Diabetes - PRS Evaluation

NOTE: 

The following files were updated on 4/9/2021 so that the output of the #feature table in the eMERGE_IV_OMOP_T2DM_PRS_algorithm script matches the data dictionary.

Files:

  • T2DM_DD_Feature_Count_OMOP_2021049.csv
  • eMERGE_IV_OMOP_T2DM_PRS_algorithm_20210409.txt
  • eMERGE_IV_OMOP_T2DM_PRS_algorithm_20210409.zip

 

 

Contact information
Contact Author: 
Authors: 
Elizabeth Karlson
Institution: 
Network Associations: 
View Phenotyping Groups: 
Owner Phenotyping Groups: 
Demographics
Age: 
 

Suggested Citation

Elizabeth Karlson. Partners Healthcare. Type 2 Diabetes - PRS Evaluation. PheKB; 2021 Available from: https://phekb.org/phenotype/1546

Comments

Submitted by Cong Liu on

Hi there,

If we can povide additional features (COD_MED_diabetes) + Common data variables asked by CC (in a certain format), do you have a script to generate the case/control you want? If yes, can you suggest what is the best way to collaborate and run the code. I will suggest you share the script which accept a standard input files (e.g. standard common data variable files collected by CC + COD_MED_diabetes) and output a PheKB compatible files.

Cong

1. What are the diagnosis codes that needs to be looked for? Also I want to confirm it’s 3 codes and not 3 visits. And 30 days or more between first and last date of any code and not same code
2. May I know what needs to be looked in clinical note (assume it’s progress note) ? Any particular term or something else. ?

The Data Floor flag is used to establish a minimal patient utilization.

It's 3 diagnosis codes from the entire ICD diagnosis catalog.  It is not specific to any disease or condition. Of patient's ICD diagnosis codes there needs to be at least 30 days between the first and last diagnosis code (there is no specific ICD diagnosis codes for this time constraint).

I do not know how notes are categorized at your site, the patient should have at least one note that is identified as a clinical note. The content of the note is not relevant.

Barbara

Is this what you're looking for in data floor? where note_ct is the count of patient clinical notes

case when (Meets_ICD9_10_Filter>0 AND note_ct >0) then 1
else 0
end Meets_DF